The Future of Construction Engineering - Fernanda Leite, Ph.D., P.E.

Felipe Engineer  0:00  
You and I met a while ago The first time I saw you, I was sitting in an audience of I think just under 700 people at the construction industry Institute, and you were about to go on stage and do a presentation on BIM and construction technology. And I was sitting next to one of one of our mutual friends, Taise. And Taise and I had been doing research for a while. And Taise elbowed me and said, pay attention, Fernando's really good, and I took an elbow right here. This is the spot where she elbowed me.

Fernanda Leite  0:32  
Does it still hurt?

Felipe Engineer  0:33  
Welcome to the EBFC Show, the easier better for construction podcast. I'm your host Felipe Engineer-Manriquez. This show is all about the business of construction. Today's episode is sponsored by Construction Accelerator.

Sponsors  0:47  
The design and construction industries come up with and build great things. But we also build in waste in how we do those things, in our interactions in our contracts in our logistics. So what does this do for our bottom line, or our next project, the best firms maximize their value by removing that waste, and only doing what's essential to the work what makes them money. Construction accelerator will train you to see the waist and give your teams the lean tools and experience to remove it immediately. All online. Construction accelerator is made up of three to nine minute videos that can be watched again and again, in the field, at the office and at home. All broken down by topic. need to learn pool planning, we have videos on the process, how to set up a room and how to kick off the team need to set up a target value delivery project. We discuss all the aspects of TVD especially costs, or maybe you just need to brush up on five as well. We have videos on that as well. You can download and print reference materials to use on site to immediately translate watching into doing subscribe today at tri ca now.com. Let's build an industry, not just a project.

Felipe Engineer  2:00  
Today's show is also sponsored by the lean construction Institute. LCI is working to lead the building industry and transforming its practices and culture. Its vision is to create a healthy and thriving industry that delivers outstanding project outcomes every time for everyone. Check the show notes for more information. Now to the show. Welcome to the show. Dr. Fernanda Lychee. Hi, Dr. Fernanda, can I just call you Fernanda?

Fernanda Leite  2:27  
Absolutely.

Felipe Engineer  2:28  
She was right. You were you were a force up there on the stage. And I was very impressed. I was. And I was like, excited about technology again, and construction, because can be a good thing. And it can be troublesome at times.

Fernanda Leite  2:40  
But presentation was part of a virtual reality study. So we worked on developing a guidebook for virtual reality for construction projects. And I think the gist of it is, we've got to be intentional about the selection of any technology that we want to pursue. And I even use the reference I think I remember using the reference of a squirrel, right thinking about technologies, a squirrel and you see something new and you're like, oh, squirrel, right. And then you go after it. And then you see another squirrel, and you go after it right. But if we keep doing that, we're not intentional and really thinking about what's the the most appropriate use case for that technology for our projects for our industry. So I think that's the if I could say what's the most important message from that it's just being intentional about the selection of technologies and using technologies and, and scanning right, you're, you're you're always looking and that's something that I like to do. I like to be scanning for new technologies and other industries and just tidbits of ideas that might inspire what what may become something ubiquitous in the construction industry in 10 2030 years. But we're always scanning I think that lifelong learning is important for everyone in our industry as well.

Felipe Engineer  3:56  
Lifelong learning is something that I absolutely love. And that's something I know that you and I share in common. We talked on the phone before we got ready for this. And I think we we touched on our just in our 15 minute phone call. I could tell that you were very much a lifelong learner. You've got the degrees to prove it in the the insatiable curiosity, I would say. And you were for those of you that didn't go to the construction industry Institute conference and saw Fernanda present. It was one of the first times hand to God that I saw a squirrel meme used in presentation. I was like, I know, I turned to tie ease and I said it's gonna be good. So, Fernanda, Please, could you tell people a little bit about yourself so they get an idea of who you are, where your passions are, and what you're up to these days.

Fernanda Leite  4:50  
So I am Fernando Lychee. I'm an associate professor in the School of Engineering at the University of Texas at Austin. So I'm part of the construction engineering. Project Management Group, which is within the Department of Civil, architectural and Environmental Engineering, I know that's long, long names. I'm also affiliated to a cross disciplinary program, which I helped create at it within civil engineering called sustainable systems. And I'm very passionate about that, because that's one of the things that when I first started at UT, and I've been at UT for 11 years, and I've I teach project management economics at the undergraduate level, I teach BIM at the undergraduate and graduate level, I teach construction safety, which I'm teaching this semester online, by the way, I can't make wait to go back to the campus into a real in person classroom, one thing that I that I desired to do when I first started at UT, is to cross disciplinary, disciplinary boundaries. And that's one of the reasons that I was so passionate about, about helping create and help make a reality this cross disciplinary program. So I led the program for a few years, and then I transitioned out of it to lead a university wide Grand Challenges initiative called planet Texas 2050. And that's a huge Grand Challenge. It's an eight year program, we have over 100 researchers involved in it. And we look at the impact of climate change and population growth in our cities, and infrastructure systems. So I do a lot of information modeling, research, virtual reality research within planet Texas 2050. But that's just the result of also my desire to really cross that silo or break down the silos. And I think that's one of the ways that we can that can lead to innovation is if we were more playful, and we interact. And we played nice with, with, with people from other disciplines. And we're always like learning from each other and trying to figure out, what can I do with computer scientists that might help me push the boundary of what I'm doing in construction engineering. So a lot of my research is sits at that, I would say it sits at the interface of Engineering and Computing. So I got my PhD from Carnegie Mellon University. And that's a very tech, it's known for its computer science school. It's known for for technology. And most of my courses in my PhD program, were in the School of computer science. So I was I was sort of trained that way coming from a program that's so highly interdisciplinary, that I wanted to continue doing that. And I still pursue that, to this day, as part of, of my role at the University of Texas, as a professor and as a mentor. And I just enjoy inspiring people, right. And it's, I'm so fortunate to be in a in a role and a job that I just get to learn new things every day and work with smart motivated people every single day. And I like to say that to my students, the hidden, they think I'm teaching them but in reality, I'm learning from them every single day. And both students in my classroom and students and in my research group, it's just amazing to work with, with each and every one of them.

Felipe Engineer  8:04  
That's phenomenal. And you put in some gems there. I knew some of that in your background. But the the one little surprise, he just caught me off guard is economics for project managers. Can you tell me more about that before we go further?

Fernanda Leite  8:16  
Yeah. So it's it's basically teaching engineering economic approaches, or techniques or tools that can help people make decisions that will enable projects to be viable. So it's simple, like net present value, return on investment, incremental analysis, understanding, finances, the basic of basics of financing, I even go through exercises, simple exercises in class to compare what the return on investment of getting a a master's degree in engineering versus only a bachelor's degree. By the way, it's totally worth it. If you look at the financial aspect only. And then we go one step further. And we say what about getting a PhD in engineering, spoiler alert, not worth it from a return on investment perspective. But so so sometimes we make decisions that are not purely financial finance driven, right. But I think that if you have the tools, you can make better decisions. So we go through other examples in class, like the real cost of owning a vehicle versus using ride sharing or public transportation, leasing versus buying and all sorts of different different problems that we solve that are relevant to the students, when at the point in time where they're, they're taking that course most students are juniors, they're thinking about the next step in their career. So they really enjoy the the course that we do in class exercise, comparing the return on investment of only getting a bachelor's degree versus getting a bachelor's and master's and they love they just laugh at the at the PhD. Piece of the analysis too. But we do that in class. As I do a lot of in class exercises. I use a technique called active learning a lot in my classroom, we learn by doing. So we do a lot of in class exercises, group discussions, mini projects, presentations, and that pretty much in every single one of my courses, I do the same thing I use active learning in all of them. And it just helps the information stick more if you're actually learning by doing than just being passive in that classroom just receiving that information

Felipe Engineer  10:22  
for those of the people watching this. And thinking how long until Philippe mentioned Scrum? Wait no more. Here it is. So in in the scrum framework for doing work. It's based on empirical process control theory, which is a lot of words just to say human beings learn by doing and that's completely aligned with with my approach and how we do things. I really like that, that you're showing those things to students early and getting their hands on. As project managers coming up in construction. Many times decisions are made on should we lease this equipment? Should we rent this equipment? Should we purchase this equipment? Should we try to own this? And those types of decisions are are not always made with, you know, thinking all the way through return on investment, sometimes people are just looking at a budget value. And do like, do I just have the money? And it's just like, if I have the money, I'll spend it to think about it in the way that you're you're showing them with, you know, economic practice, you know, like, should I buy this car should at least this car, how long am I going to have this car? What kind? What types of needs? Do I have just a little bit more of those questions, I think makes for much stronger project management, and better financial decisions and outcomes on site. So super love that you're doing that. So thank you for that. Appreciate that a lot. You mentioned that you went to Carnegie, I did a little digging. And I like I followed some of the work that Andrew has done, you know, back in the day, and he had the slogan, I think he says My hearts in the word

Fernanda Leite  11:42  
I lost count of how many times would you hear that? And and I think that's the philosophy of the university. But I think it goes that that permeates in a lot of different schools. It's especially engineering like we people are very hard working. And and I interacted mostly with graduate students at the time I was a teaching assistant. So I interacted with some undergrads to at Carnegie Mellon. But people are just extremely hard working in general. And I think starting a university and creating that that workforce there. So it started as the Carnegie Institute of Technology. And then it merged with the Mellon College of Science, it really was creating a workforce for the future that was needed for that region. in Pittsburgh, Pennsylvania, that really had a was a leader, world leader in steel production. So engineering was a big deal. At the time when Carnegie Mellon was created. And now they sort of shifted, it's a tech hub. Now, and much like Austin is if people say Austin is the new Silicon Valley, we have a lot of people from California. Moving to Austin, a big part of this attraction is when you have that critical mass, when you have people being trained that you can then recruit these, these big companies that are moving to those cities can recruit from as educators we feel we have a huge role to play in the workforce to have training those next generation leaders of of training people that are going to change the world. And now I'm going to make a plug for the University of Texas slogan, which is what what starts here changes the world, we're making a mark, and we're trying to make the world just a little bit better.

Felipe Engineer  13:18  
I'll give you a double plug on that. I was at a job and the project manager had gone to school in Oklahoma, which is some people say is North Texas. I just like doing that because it's fun. But he mentioned University of Texas in particular said I went here, but this is a really good school to and they teach these types of things that you know, all the schools in this area kind of trying to model themselves after University of Texas. And this was just a casual conversation that he and I were having because he was talking about things like net present value, and some of those same economic principles. And that's what just jogged my memory and made me think about Kenny, so thank you, Kenny for giving a shout out to University of Texas, which didn't even go to so I can tell you know from those things that you are very passionate about this, it comes across and I knew you'd be an awesome guest. Just for that alone. What got you started in construction like where, where this come from?

Fernanda Leite  14:17  
I'm originally from Brazil. So I grew up partly in the US partly in Brazil, because my father went to Texas a&m University for his PhD. He's an agricultural engineer. he actually got his Master's and PhD there. So we lived in College Station while I was growing up for approximately seven years between his Master's and PhD living that university I spent, I would spend a lot of time at the university in my summers I would hang out in while he was working on his experiments in the lab, I would hang out, wait for him and walk into and spend a lot of time in the library in the main library on campus. And then he would pick me up at the end of the day and the library would go play tennis together. So I saw what that life of being a researcher was like through my father and He's a professor now. And in Brazil, when we went back to Brazil, I was 14 when he's finished his PhD, and my grandfather, his father was a developer of high rise residential commercial construction in our hometown, in the Northeast of Brazil. I remember going to one of his job sites when I was eight years old, seeing the chaos and just feeling well, this is really cool, right? This is there's all this stuff going on all these people working here, this is such a dynamic environment, I want to do this, I was in college. So fast forward, I know I'm an undergrad in college. And then I started since I could speak English fluently, I went back to Brazil when I was 14. So I finished school and did my undergrad, master's and came back to the US for my PhD. But when I was an undergrad in Brazil, I taught English in an after school program just to make money, it was better money than flipping burgers as an undergrad. So I say, I enjoyed, I enjoyed that so much. I just enjoyed teaching, and seeing people learn and seeing the light bulb moments in people. And I remember having a conversation with my dad saying I really enjoy teaching. This is something that I'm passionate about this, but I'm also passionate about construction. So I'm a little bit conflicted. I don't know what to do moving forward in my career, and my dad said, well, it's obvious, you need to be a university professor. So just put those two passions together. But in order to do that, you're going to need to get a master's in a PhD. Okay, so like, I think I can do that. So then I did undergraduate research. To prepare for for a Master's I, I went to probably the the most prestigious Master's graduate program in construction engineering in Brazil, which is in the south of Brazil, in the city of Porto Alegre in the Federal University of Houston to the, to the long name all the universities that the top universities in Brazil are Federal University and the name of the state. And that's where vice and I meet, we actually never, never, were in the same. We weren't there at the same time. So she finished her master's, and then went to Berkeley. And this is tyese LVC. She's an associate professor at San Diego State University. And she's a very good friend of mine. And she also got her master's in the same program that I did in southern Brazil. And we've we've been friends, since we crossed paths and academic conferences, and every time one of us visits each other's cities for work or, or or for for fun, we definitely always find find time to hang out together for being there. And that program was, was just a wonderful experience for me, I got to focus on developing myself as a researcher, I had a wonderful advisor, Carlos Formosa, who was the same advisor advisor that that he had for for her master's program. And when I was there in the program, there's a professor from Carnegie Mellon, who's Brazilian who also got his master's from that same program with the same advisor that I use, and I had, and his name is Lucy seidelman. And now he's a professor and department head at, in the Civil and Environmental Engineering Department at the University of Southern California, in Los Angeles. And he would go every summer and teach a one week course on machine learning, applied to civil engineering. And I was like, well, I'll take this course and sounds fun. And I had zero background in computing at the time, right? Zero, I take this course. And I'm like, wow, this is like super cool. I was like, Mind blown with with the potential of using like machine learning. That's really cool. And, and at the time, I think he called it data mining, but still, you know, I was really passionate about using How can we leverage data to make more intelligent decisions. And at the end of that one week course, he actually asked me, Hey, have you thought about getting a PhD? And I was like, yeah, I'm planning on getting a PhD. But at the time, I didn't even consider going to the US for my PhD. I was I was thinking, Oh, I'll just stay here, continue my PhD and, and get a faculty position in Brazil. That was my, my thought process at the time. It's like, Well, why don't you go to Carnegie Mellon? Well, I hadn't thought about that. That sounds really cool. Then I the only thing is that well, I'm married. My husband is also a master student. Here we were. We got married and started our master's program together a month after we got married. So we're both from the same hometown. We went to the same university med school we and my husband is is the Associate Director of Research at the construction industry Institute, Daniel, Ollie beta. We don't have the same last names which till today I hear about 18 years after we've been here, man married I still hear about that. But I joke with them saying that the year we got married was the year that Brazil changed its law so the husband could change he could have changed his last name to my last name, if he had chosen to so I'm like, okay.

Felipe Engineer  20:00  
Fernanda, my last name and my wife were hyphenated, the same.

Fernanda Leite  20:04  
There we go.

Felipe Engineer  20:07  
Could use me as an example? Take that Daniel! Yeah, take that. I've known Daniel for a couple of years. He's a good sport.

Fernanda Leite  20:15  
And so then then I was like, Okay, well, we have to find a way to for the two of us. We always really prioritized both of our careers. Since we were undergrads when we were trying to decide where to go for our masters. It was always a matter of where can we both find opportunities. And we both managed to get into Carnegie Mellon. And we both managed to get graduate research assistants. So working as a PhD students, that would be fully funded PhD students that in exchange for working in research projects, we would get, you know, our tuitions, we'd get a stipend and whatnot, which made it feasible for us at the time. Carnegie Mellon is a private school and tuition is, is pretty, insanely high. So if it weren't for that we wouldn't have that have had that opportunity. But Fun fact is I was also admitted to UT Austin. So the two programs that I had my eye on and when I was applying for a PhD were Carnegie Mellon and the University of Texas at Austin. And my colleagues at UT always joke saying, well, you ended up here anyway, we ended up getting you and, and so then we went to Carnegie Mellon. And that's and I remember my first semester at Carnegie Mellon and one of my committee members, Omar Aiken, he was one of one of the first PhD students of the biggest name in bam, a chuck Eastman. So Chuck VSAN, passed away last year he, he was one of the authors of the BIM handbook. And he was a professor at Georgia Tech. But he also was a professor at Carnegie Mellon in the 1970s. I believe. I'm actually in the leading one of the tracks for the Eastman's symposium that will take place on May 13, in honor of his his legacy, and academia, Omar, so I was meeting with Omer one day, and he and I was taking this course called data structures. And it was probably the hardest course that I had ever taken. I this is a person that didn't have computing background. And I was having to code an airport simulator. And so I was having to learn that and I had never felt that kind of challenge in my life. And so I remember calling my dad and telling them, you know, I think they made a mistake, they shouldn't have admitted me to this program. I don't think I deserve to be here. And so I went to Omar's office and I and I told them, and he was like this quintessential academic, he was drinking sipping his his tea and I walk into his office, he puts his t down.

Felipe Engineer  22:43  
I'll tell him.

Fernanda Leite  22:44  
Yeah, there you go. And I tell him that look, if what it takes to succeed in this program is to finish this course successfully. I don't think I can do it. This is not for me. And then he puts his t down. And he says, you know, I've, I've seen students like you before, you're the type of student that everything just came naturally, and you learn really easily. And now you're you're finding the one of the biggest challenges that you've faced in academic career. One thing I can say is that, you know, there's hundreds of you. Here at Carnegie, you are always like the standout in your academic work. Now, there's like hundreds of you, you go back to your office, and you work hard. And that's again, what the whole philosophy right at Carnegie Mellon there is, is in the work, and go back to your office and just work hard, and you'll do it, you can do it. And till this day, that was the proudest A minus that I ever got. And, you know, all of my degrees, I worked very hard, I spent about 40 hours, but I finished that airport simulator using queueing theory. And I was done. And I come and that really empowered me to really push. And that's a lot of what what a Ph. D. program is it really pushes you to your limit. And you really have to feel feel comfortable with that feeling of possibly things going wrong. And that's that also is what research is like. So if we knew what the outcome of research would be, it would be called research. And I always tell that to my students, like research is about failing, you learn to you become a stronger researcher. When you learn from from the failures when you learn from those mistakes. You know, if you see somebody's curriculum, you only see the winds, right? You only see like, if you look at my CV, my students are like, Whoa, look at all those publications. Yeah, but you don't see all the rejections Right. Right. And there's plenty of them in our career, but we learn from them, we just become stronger people. And that goes for any profession, right? We all learn from our mistakes and that just makes us stronger, but being comfortable with failure. It's because you're always pushing that boundary and that's something that you you learn to be okay with as you progress in your career.

Felipe Engineer  25:00  
Man, I love it that Omar pushed you That was really cool pushed you. And then you honed in on that perfect Pittsburgh accent. I'm still trying to hear any hint of Texas and nothing's coming through.

Fernanda Leite  25:10  
People have a have trouble figuring out where I'm from, even when I speak Portuguese. So when I go to Brazil, and I give talks, I one of the questions that I always get is where are you from in Brazil? So I don't have like a foreigners accent in Portuguese. But since I moved around so much, I have like that standard Portuguese accent.

Felipe Engineer  25:31  
The Rosetta Stone textbook Portuguese, right?

Fernanda Leite  25:34  
Yeah, yeah. But if you can tell I'm Brazilian talk to my husband, Daniel. He has a Northeastern accent. DICE has a very Northeastern accent. But mine is more like standard. So I always get that question like, Where are you from. And people are surprised when I say I'm from here, this is my hometown. Like what?

Felipe Engineer  25:56  
It's got to be cool to be in your class too. If you if you teach like this, oh, you've got to have all your classes just filling up right away.

Fernanda Leite  26:02  
For the 11 years I've been teaching BIM, oftentimes the waitlist is larger than the number of seats in the class. So usually, if I have 36 seats, the waitlist is maybe 50 students long. And so so people enroll in that class, like in the first seconds that enrollment opens may enroll. And you can I can see that from the timestamps of when people enroll in the class. It's literally like seconds apart.

Felipe Engineer  26:26  
And then it fills up a lot of people saw them come out in the 90s. And they're like this is going to revolutionize construction. revolutionize and that's what people were saying and and we still see that there's a lot of companies latched on to it, and there's still many organizations, I honestly get contacted by BIM startups, and and even well established companies on LinkedIn weekly. What got that sparked out for you? And how do you see the potential.

Fernanda Leite  26:53  
So when I was a graduate student at Carnegie Mellon, there was a building under construction, The New School of computer science building in the middle of campus, and that was called the gates building. It was funded by Bill Gates, the general contractor for that project was had heard of this thing called BIM. This was in the early 2000s. So he had heard, they had heard about this thing called BIM and my PhD supervisor, Virtua kinji, she had a good relationship with with this company, she had had several students do studies in this company, that's something that we typically do in the construction engineering domain is our lab, our active job sites, active construction projects, so we send our students to to a lot of projects, and she did the same at the time. So she had a really strong relationship with that company, she sent me to that job site, learn from it and do different things in that job site. And one thing that that sparked my interest was design coordination. And then with this thing, this new BIM thing starting, I went my recommended to the general contractor, the GC had actually hired a third party company to create the model out of the drawings, right, the 2d drawings of the project. And there was this BIM model, it had architectural structural, mechanical, electrical plumbing in the model, and I was like, I saw this model, and then they were about to start design coordination. And then I go to them, and I say, how about we do design coordination using the model, as opposed to using 2d drawings, overlaying them on the light table? Oh, my gosh, it was like, major pushback from the subcontractors and they're like, weird, didn't sign up for this, we were hired. And then we're not going to do this, we're just going to keep doing business as usual. So it took them was like something like seven to nine months to coordinate that project. 210,000 square feet, because they were overlaying the 2d drawings on a light table, I participated in every single one of those weekly coordination meetings, that lasted anywhere between five to eight hours, each.

Felipe Engineer  28:50  
Pause for everybody, just to give context, this is the 2000s. This is the early 2000s. I mean, BIM is arguably a little bit early 90s, but like 1995 was kind of mainstream and anyone can buy, I mean, not cheaply, but you can buy this type of software package and get a really good computer. And, and you can use it, and were there were some very forward thinking companies in the 2000s using it. So I just wanted to just remind everybody what the time period is, because a lot of people just, you know, take it for granted now, because it's just so every day on sites, kind of like safety glasses now.

Fernanda Leite  29:26  
Ate out of me every week. And what I would do in that meeting, as I would document what, who, who were the trades that were coordinating at the time, what were their information exchanges. So what questions were they asking each other? So this is 2d, so a lot of the questions was, what's the elevation? What's the clearance here? So a lot of information that in 3d would be pretty easy to you wouldn't even need to ask that kind of question. I document all of their information exchanges and all of the clashes, the types of clashes, and the number clashes that they found for every single one of those meetings. And then I would go back to my office at Carnegie Mellon at the end of the day and repeat that same section of coordination with the same trades that work for those five to eight hours. Repeat that and do the same thing in a couple of minutes. And then document what I found how many clashes and the types of clashes that I found, we have a huge challenge in in construction engineering, as our projects and from a research perspective is our projects are unique. So they're like one off prototypes, we have one project. And that's it, if you're trying to do a study that you're trying to create, like a control group and a treatment group. So the treatment group would be using BIM. And the control group is the current approach. So 2d on the light table, that's really hard to do in in our replicant to try to replicate that or create that environment in research, right. So I really had that opportunity. in that study, I had the BIM model, I had the skills to be able to update it, if there were any changes based on on the coordination that was happening. So I did that comparison, what I found was that in the the 2d approach they had, so I use two metrics from so this is now my professor side. So two metrics of from the information retrieval, domain, precision and recall. And I'll use the same explanation that I use in my BIM class. So think about Google. So when you Google something, let's say the University of Texas at Austin, the first result in your Google search will likely be you texas.edu, the main UT website, that means Google has a high precision rate it returned. And then the top result, what you are looking for, probably if you just Google the generic title of the university, you're probably looking for the main website. But if you scroll all the way down to the results page, you're going to see that there's hundreds 1000s of pages that you can, that that Google found that cite the University of Texas at Austin, that means Google has a high recall rate it's catching. It's finding every single site, web page that cites that term that you were looking for. Now, going back to to design coordination, how does that relate? So if you look at expert coordinators, or subcontractors, subject matter experts, when they're doing a coordination in in 2d on a light table, they have a very high precision rate. So the clashes that they found were real clashes, that that really had to be addressed. But they had a extremely low recall rate, something like 10% 15%. What that means is that they weren't finding nearly as many clashes as they should. And that resulted in field detected clashes. And you know what that leads to change orders, change orders, yes, and delays and increased costs and all that. So that's the, that's the problem there with that manual design coordination. In the BIM world, we get very high recall rate. So basically, so long as it's modeled, we'll find it. But we're currently getting lower precision rates, which means we get also get a lot of noise with our data. So we get a lot of false positives, and that has to do with model quality. So sometimes, if you, if multiple people are using different software systems, you export to a model coordination software, some objects may explode, like an example would be a valve modeled after actually about 15 different objects that compose that Valve object. And if it's clashing with, let's say, a section of a duct, that will return 15 clashes when in reality, only one is a true positive clash, and the rest are false positives. So noise in your data, so you've got to clean out that's what the VDC engineer or the BIM coordinator is going to do. So they sort of figure out what are the clashes that are real clashes, they Let's clean out the noise, and they figure out approaches to clean the model and you run through just the clashes that need to be solved with that group of people. And and that now we see these coordination meetings happening in an hour zoom meeting, right? It's a lot more efficient. Now, just because we know we get high recall rates, so long as the information is in the model. It will catch it. And and BIM coordinators are now more more efficient to and how they prepare for those coordination sessions.

Felipe Engineer  34:21  
What was the difference between the 2d light table and I know it was it was a while ago versus what you were seeing what you're capturing and doing unless the time.

Fernanda Leite  34:29  
Yeah, so it's that was the precision and the recall rate. So in the 2d light table, we were getting varies to by by the pairwise comparison that you're doing by what trades you're comparing, but I would say roughly about 20% recall rate between 10 and 20% recall rate, so only finding 10 to 20% of the actual clashes that you should be finding the precision rate in the 2d light table was nearly 100%. And then we had the opposite in in BIM so they it's an inverse relationship. Nearly 100% recall rate, but because of all the noise in the data, the precision rate was anywhere between 20 to 40%. So the message is if we want to try to decrease that noise in the data, so it's important to really consider the quality of or your model how much information you want to put in your model that's relevant for that design coordination. So one, one message that I always tell my students in classes, just because we can model doorknobs, or we can model all sorts of little details in the model is that even relevant or that process that you're trying to, to use that model for for design coordination, you're probably not going to need that information. So you shouldn't be modeling it, because it's going to lead to, you know, more noise in your model. In the design coordination process, it's more information for you to keep track of and having to manage when there's updates in the model.

Felipe Engineer  35:52  
I love that practical spin on there, I'm gonna replay that that's gonna be my clip, to all the people that want to model everything and have a digital twin like, no, it shouldn't be a digital twin. Exactly. It should be a digital practical twin.

Fernanda Leite  36:06  
Yeah. And it also depends on what you're using the model for, right? So you can't have a super detailed model, maybe you, you want to test out the connections in the exterior enclosure, or every single piece of part or a specific sequence of activities. But in a small piece of the model, you have a 210,000 square foot building, you probably don't want to have all that detail in it. So it's just gonna be a lot of time invested in modeling for little to no return on that investment.

Felipe Engineer  36:37  
There's that economic thinking, there we go. Yeah, so practical, like what's gonna be the return? Like, we only have limited time? Absolutely. Yeah, some people like Bill Gates have nearly unlimited resources, but many of us, but even he is subject to limited time.

Fernanda Leite  36:51  
We all only have 24 hours in the day. So we have to be very effective in how we use that time, so that we're solving the problems we need to be solving.

Felipe Engineer  36:58  
Right. And on that note, I could tell that, you know, often I hear from from people that come on the show that are lifelong learners, and definitely having an impact. And I can say 100%, from what I've heard, the last time we talked when I saw you on stage, and now you're absolutely having an impact, a positive impact on our industry. So thank you for that. If your students, they'll say thank you enough on behalf of all of your students, and I wasn't one. But thank you for your contributions to our industry and profession. I do want to ask you your opinion on continuous improvement. What do you know about it? Because I mean, your friends will tell you so you know that lien is gonna come up at least every other time you talk. So what do you what do you think about that?

Fernanda Leite  37:35  
Right? So so I think that we can always improve things I think it's a matter of, and also you can only improve what you measure, right? So we understand a problem, any problem. And we understand, we can we can measure a process or a time or whatever metric it may be, we can figure out ways to improve it. But we can only improve what we can measure. But I think having that mentality of that open open mindedness of saying, We anything, any process can can be more efficient can or more effective, right, any single process, I think that's the philosophy that we have to carry out and everything that we do every day. So we can always improve. It's a matter of and I tell that to my to my PhD students as well, when they're writing papers that sometimes you you want to refine that eternally, right. So you can you can be working on the paper and never be done or a PhD dissertation never be done. But at some point, you've got to say it's good enough, and let's move move forward. But understanding that there's always an opportunity for, for improvement.

Felipe Engineer  38:47  
Perfect definition of continuous improvement. Thank you for Nanda, the potential for BIM today. Now it's 2021. And computers are much more powerful. I think my my phone computer is more powerful than the laptop I had when I first got into the industry back in the 90s. Where do you see things going now with with BIM and computing getting smaller and faster.

Fernanda Leite  39:07  
We can now do more and maybe more intelligent, automated decision making. So design coordination as an example, even though you're using a model coordination software, a lot of them are replicating what we do manually, they still do pairwise comparison between two trades, that's basically replicating what we do by overlaying 2d drawings on a light table. But we're using a machine to automatically find those clashes. And a lot of that can be automated as well. And route planning can be automated. There's a lot of research going on in generative design that can help automate, imagine like an autocomplete for for design coordination, right. So I think that we can leverage more now that we have more data, leverage more artificial intelligence to automate a lot of processes just to make our work more efficient. Think about you know, 10 years ago or or no more than that. 1520 years ago, we were doing the coordination and and five to eight hours meetings now it's a one hour meeting, maybe 10 years from now, it'll be a five minute check in, right. And I think that with automation comes these efficiencies in our processes. And we can, then As humans, we can dedicate our time to more high level thinking, strategic thinking, right, because a lot of very repetitive tasks. And that's what machines are great at machines are really great at doing very repetitive work that us humans were not good at, we get bored, we need novelty. And, and I think we should leverage technology to do all the repetitive tasks, so that we can dedicate our time to think about the construction engineering aspect, really the complex decision making, that is really hard for a machine to do and requires a lot of computing power. But for us, it may be easy. There's a whole field in computer science called human computation. And an example of this field is you probably when you were buying like, back in the pre pandemic days, let's say a ticket for a concert or a sports event. And when you're buying the ticket, you see those, those squiggly words, you know, those images, distorted images. So so that's a capture. Our, that was a startup from from years ago, it was basically to try to figure out if it was a human that was buying that ticket, or a bot that was then developed by scalpers, right, so they could resell those tickets. So it was checking, because that computer vision 10 years ago was a very challenging task. So if you look at a distorted image, or distorted word, that's as a human, you can quickly interpret that image those distorted words. But for machine, that's a very, very difficult task. With that in mind, then we can use that to check if whoever's buying that basketball ticket or that concert ticket is a human or machine, let's just show distorted image of a word. And if they get it right, then it's a human, then the researchers that develop that created a new approach that well, we were not getting anything out of this where humans have this computation power that to do something to interpret these, these distorted images that is so hard for machine to do. But we're not leveraging this, we're just confirming the words that we already know, well, how about we start, we start using that power for good. So basically, that's where reCAPTCHA CAPTCHA came in, that you now see two images, right? One is, is a distorted word that checks if if you get it, get it right or not, if you're a human or not. And then the second one is a new word. So if you got it right, then the likelihood of you getting the second one right is high. But that second one, you're actually helping populate a database that's helping decode scanned images of old books that might be hard for like an automated, you know, coding to find because the words were distorted, but you're helping do that until that, that...

Felipe Engineer  42:46  
Collective computing.

Fernanda Leite  42:47  
Collective computing! And that's because that's the power of human computation, there are tasks that are still very, very difficult for machines to do, and that require a lot of computation that we can leverage our brains to do. That's the same idea that when I started my National Science Foundation project that I like to call living BIM, and living BIM is, is when I was writing this proposal, I thought, well, buildings, facilities, they're like living organisms, they change throughout their life. And we have all this this cool, cool stuff that's going on in VDC and bam, in design and construction, when when General Contractors finish the project handed off to the owners. I wasn't seeing like owners using that model for much, right? How can we get people to actually use these these models in the lifecycle of the facilities a longest phase of the project lifecycle at UT, we build for 100 year construction 100 years, we should be able to use these models, and we're not and why is that because current state, we don't have the in house capacity in owner organizations to update these models. And so the whole idea behind living BIM is well, how do we remove the human from the equation? So again, that human computation aspect, how do we get humans to focus on what they're good at? And what can we automate? Can we automate the process of updating BIM models without a human having to open a model authoring software system? And let's say moving a door from one place to another, how can we automate that and in this project, and the route that we ended up taking was using a combination of computer vision and deep learning, which is a subfield in machine learning, we ended up using an approach called transfer learning a lot of databases that exist computer vision database that are publicly available image net is one and a cocoa is another So Microsoft, cocoa is another you have a lot of these images that are labeled, so it's like teaching a kid right. So my when my daughter was was, you know, a baby, I would point to something, you know, a mug, or a cup or, or, or pen, whatever it may be. That's how kids learn. And that's how you can train algorithms to learn what what the These objects in the real world are as well. But we need lots of examples. So we use transfer learning, because there's a lot of our data labeled out there already in the built environment, like for walls for Windows for, for objects that that any human would know what, what they are. But we augmented that that database with our own data that we collected, we manually labeled with the help of dozens of undergraduate students, a very hardworking undergraduate student.

Felipe Engineer  45:26  
Undergrads, they make research possible. 

Fernanda Leite  45:27  
And so there you go, they definitely do. And I'm very, very lucky to work with super smart undergrads, they labeled a data that was in our specialty data set. So data set that contains objects that your average person would likely not know what they are, but they are part of building systems, things like variable air volume box, or an air diffuser, or a sprinkler head. So things that are a little bit more specialty, and in the built environment. So we call that 3d facilities. And it's publicly available to so that it's labeled data of anybody can can use that data. And we publish that via zenodo, which is, you know, a website that people share code in, then we use this the database that already existed our data on top of that, we use that to train a deep learning algorithm to learn what they're seeing in the built environment. And if they know what they're seeing in their built, built environment. And if we show them an original BIM model, we overlap, let's say, a point cloud with a BIM model. And there's a difference, they know what that object is. And then we can use generative design to actually move that door to where it should be the physical world. So that's where we're at right now we're in the last phase of this project that we're actually in that the modifying the model face, we did all the training, we built the database. Now we're in the the generative design piece of it.

Felipe Engineer  46:46  
That's really cutting edge.

Fernanda Leite  46:47  
Yeah. And it's fun. It's it's extremely fun. I think that the two PhD students that worked in this project, one is now a professor at Arizona State University, Thomas tarnowski. And the other john longmire, they both had to take a lot of courses in computer science, they had committee members, or have committee members members from the computer science department. So they're another example of crossing those disciplinary boundaries, we can only do that this this kind of cutting edge research, if you have the skill set. So you sort of have to go and learn from computer scientists about how to optimize algorithms. And we're lucky to because UT has the largest supercomputer of any university in the world. This is our tac, our Texas Advanced Computing Center, we're able to tap into tack and run these computations, which is a lot quicker when we're training the neural network. It's it's a lot of data. So we can do that a lot quicker if we're using a supercomputer as opposed to a high end office machine, right, that desktop computer or high end laptop. And then there's several different approaches that they they're working on, john is working on developing, how do you parallelize or you leverage parallel computing to really optimize those processes. That's how we got to when you pointed to your phone, right? How we got to the computing power, it's a lot faster than my computer when I started working in this field, that's called Moore's Law, right? Every two years, computing power doubles, right. And so it's, it's a matter of multiple people also figuring out ways of, of how algorithms can work more efficiently, we really need to learn from a lot about computer science, from computing to be able to push that boundary and enable our industry to process information more effectively. By really leveraging the power of computing and artificial intelligence, we can do more we're all collecting now that we're all using all sorts of data management systems, project management systems. In our projects, we have lots and lots of data. We're drowning in data, we need to make sense of data. And we need to do that, by again, leveraging what machines are great at, which is doing a lot of repetitive tasks and doing a lot of automation, artificial intelligence. And that frees up time for us humans to be doing human computation. But the high level really complex tasks are hard for machines to do.

Felipe Engineer  49:00  
And we do borrow quite a bit from computer science like even and again, here's another reference to Scrum. Jeff Sutherland introduced me to Brooks law. And Brooks law is something that came from computer science for specifically computer programming projects, adding more people to a project that's late makes it later and they had data from software programs because they could pull data out of the coding programs to see what timestamps and how things had done. And when projects got behind schedule, human instinct is just add more people to it, which works when things are simple, but in programming, it's complex because it's a language and things when you change one thing it has multiple step consequences that are hard for us to see immediately crossing of all these disciplines goes back to that we started with empiricism learning by doing your researchers learning had to learn new things you're they had to bring on committee people where they found gaps. I'm really looking forward to the day where we can have a living model and because we do a lot of tenant improvements and changes and now has a client said, Oh, and by the way, here's an as built model of the building right now we don't we're lucky if we can get as Bill 2d drawings of what was done. And we've spend enormous amounts of human effort just uncovering What's there in front of us. Whereas if we had that living model, like you're working on, then, you know, people can get focused on what they really want, which is to have a building that works for them suits, their needs fit for purpose.

Fernanda Leite  50:22  
And that's all about figuring out what are the processes in the AC industry, that could be made more efficient, it's again, going back to the conversation that we were having about continuous improvement, there's always things that we can improve, we just have to find that, that nugget. And that's another thing that we I always tell my students and and this goes again, with that ci presentation on virtual reality. Some people like want to join my research group, because they want to do research in virtual design and construction or BIM I want to do research in bam, well, what is the problem you're trying to solve? That's the first asking, right, that's what it might not be been right? Something sometimes pen and paper might be enough for the problem you're trying to address. But I think that it's and that's the challenge with technology. And that goes for industry that goes for research track idemia, it's the hammer in the nail problem, right? So you have this hammer, whatever technology does your you have, everything starts looking like like a nail, you can break everything around you. So what you really need to find is that nail, so what is the problem you're trying to solve? What is that process that you're trying to make more efficient? How do we measure efficiency in that process that you're trying to do? And then you start trying to try to figure out what is the most appropriate technology to solve that problem? Does that exist? Does that exist in our industry? No. Does that exist in other industries? Something that we can adapt to? Aren't those? No? Well, maybe we've got to come up with something that's also that where we came with the inspiration for one more recent CIA project that I led called path to the future and the path to the future. The whole goal of the project was to inspires step change and innovation in the construction industry. The whole inspiration in this this project, what came from the Jetsons cartoon. So think about it. So the Jetsons that was developed in the 1960s. And a lot of the technologies, I watched that I watched the real ones when I was a kid. So in the early 80s, that 20 years before that was about it was still really cool and innovative. And a lot of those technologies like wearable watches, flat screen TVs, flying things like drones, jet packs, those inspired an entire generation of engineers, and scientists actually make those things a reality. So sometimes we really have to be a little more playful and looking at trying to identify what is the problem that I'm trying to address? What if I don't have any constraints? What if I can just create something totally crazy and totally new, just like cartoonist did in the 1960s with the Jetsons. That's what inspired us in path to the future. And it was super fun to work with this group, we developed basically a series of workshops. The challenge with industry folks is that they're always firefighting, it's hard for them to put themselves in the mindset of being creative and being playful because they're, they want to solve a problem and move on. And they always have that previous problem lingering over them and impacting the kinds of solutions they make. So how do we then develop a creative process with people that might be unencumbered by previous work experience, that's again, where undergrads come in, held it in three different universities across the US one at the University of Texas at Austin, one at Carnegie Mellon University, and one at Georgia Tech. And we had over 100 students participate in that show up in a Saturday afternoon. And the way that we convinced them to show up, we just put posters all around these three universities, is, we call that Mars industries. So come and help us build the next generation habitat in Mars. And so people signed up. And the reason why we chose Mars is because we need to put ourselves outside of our boundaries, our constraints are typical access to resources that we may have. If we put yourself in a totally different environment, you're going to have to be creative, because you don't have access to the same resources that you have, let's say here on Earth, Earth to build something basically gave them a series of tasks, many tasks throughout the day to solve that problem. To build a colony on Mars. They came up with literally 1000s of ideas, and we collected all the post IDs that we have to then organize and make sense of I had posted nightmares after those workshops because all I could see were all those posts. That's fine falling off the wall.

Felipe Engineer  54:36  
Like just...

Fernanda Leite  54:38  
Yeah, crazy. No, and it was like 1000s facts and then we made sense of all that information, took them to SMEs in the industry. So this was the technology and innovation committee at ci they sort of made sense of all that information, identify 12 technology enablers that would again remove burdens from the heat So we can focus on high level thinking, what can we do? What are these technology enablers that can help us focus on what we are really good at things like modularization, like automation are examples of a technology enabler. And then we we created a program, we thought that was really fun with students, but then how do we change that culture in our industry? So we created an ideation program called challenge teams, the way that it works is we put together a group of people from industry from different companies, and they work remotely and they work on solving a problem, one problem that's given to them and they have six weeks to do it, they have to identify technological solutions for that problem, thinking about three different timeframes in mind, what if you only have five years? What if you have 10 years? What if you have 30 years 510 30 basically thinking about what technologies do we have now available that we could help help solve this problem? What technologies other industries may be developing? Or that we can then adapt to our industry? And what are those crazy ideas that a cartoonist from Jetsons might think of applied to our industry, putting them in that frame of mind, and then the problems that we give them are extreme. And what I mean by extreme is, that's again, trying to replicate the idea of putting students in Mars, giving them a challenge that seems so crazy right now that they're, they're really going to have to think outside the box. An example is how do we build a project with no humans in a job site? I mean, zero humans, it's an extreme problem by design, you really have to make people think, no, it's I don't, I don't mean half the humans that you have. No, it's zero. And when we give them those challenges, they always look at us like, really? How are we going to come up with that you have six weeks, come back and tell me in six weeks, we just have so much fun in this process. So we're already in I think, challenge teams. Well, we launch two challenge teams every few months. And it's all managed through through cis technology and innovation committee. And I just launched the teams and I'm brainstorm some of the challenges via via the committee that leads this process with a lot of SMEs from the technology and innovation committee. But it's just super fun for me to be part of, and to hear what people come up with and the inspirations that led to their solutions. They all they like to engage their kids drawing cartoons. Some use Hollywood as as inspiration, hollywood movies, action movies. So it's just fun to watch, you know them having fun in this process and the ideas that they come up with. And it's a culture change, right? The whole point is how do we create that culture of continuous improvement of really trying to think outside the box of removing yourself from your day to day firefighting mode, and giving yourself that that opportunity to be playful? And to think about things from a different perspective, that'll help change that culture in our industry and inspire our industry to lead that that step change in technological development?

Felipe Engineer  57:52  
I think it will, you have a great framework there for people to completely step out the magic. And that is how you created some constraints, even though you said like, constraint free, you still have the constraint of like how much time you have, and then you've constrained them on the thinking, I think it's brilliant, because things that I can do right now I'm gonna, as a human being, I'm serving the landscape of what's possible today. And then I got to think, completely different, like, just let my mind wander to go to that 30 year mark, if you think backwards, 30 years ago, you know, we didn't give children cell phones. And now every kid in most of the industrial countries are born and given phones immediately to pass the time with radically different how much closer technologies come to us compared to what it used to be.

Fernanda Leite  58:33  
If you think about technological development rate of development is exponential. So we all know, what an exponential curve looks like now with COVID. Right? So and it's the same thing with with technology in our industry, when when did we start seeing smartphones not too long ago, and and it's no iPads, you know, that's, that really has revolutionized how we interact with data and how we process information and how we access information. It's, it's really impacted so many industries, just that availability of data that we all have now and all sorts of different types of fields.

Felipe Engineer  59:09  
You took me out beyond the edge of where I thought you're going to take me with, with technology. So thank you for that. Is there anything in like in the stories or in the papers he written or the research, you've been a part of? That's really stood out to you as just a fundamental success that if more people knew about, you think it would make the speed the change up in the industry even further?

Fernanda Leite  59:31  
That's a really good question. I'll go back to the human computation aspect. I think we do have to always ask ourselves, what can we do better? And is there technology available out there for that, for example, virtual reality might work really well, for some use cases, like training, it might not be great for other use cases that maybe augmented reality or mixed reality might work better. And so I think always asking yourself, what's the best technology for that use case that problem that you're trying To solve isn't available, can we adapt it from another industry? Or do we have the capacity to develop something new? If that doesn't exist? I think understanding that the process that you're trying to improve is the first step. And for technologies, I would say it's the next step, because likely, there's something that can help you be more efficient. And it might not be the perfect tool. But it might help you think as you're doing right and develop those those ideas as you're tinkering, trying to make that thing work, and help you think out loud as you're as you're doing. So that learning by doing as well.

Felipe Engineer  1:00:32  
So I like that, do you have like some secret crush on industrial engineering? Because a lot of the stuff that you're talking about? Or it seems like you're very attuned with a lot of the folks that I know that are industrial engineers, I mean, just keenly process oriented.

Fernanda Leite  1:00:47  
Yeah, I think I think that's fine, lean construction background, because that's the master's program that I went through in, in southern Brazil worth, it just came out of that's very lean, focused. And, and I think that's definitely something that that sticks with me is is looking at how can we map processes? value chain mapping, right? How can we identify what's non value adding and remove that from the equation, but then I, I like to think about it from a computing perspective. So if we understand the problem, what pieces and parts can be automated in that in that process? What's the fact that we can remove from from that process, or maybe completely change it.

Felipe Engineer  1:01:24  
So you get more what you want faster, and with less effort.

Fernanda Leite  1:01:27  
And hopefully, we'll have like a five minute work day in the future. We can all be hiking all day.

Felipe Engineer  1:01:35  
All day, we can just be taking things in, we could be learning how to program more.

Fernanda Leite  1:01:41  
Because we still need humans. It's not like we're saying a lot of people are afraid of technology or afraid of automation, all of that. But what people are going to lose their jobs. No, it's just going to change the kinds of jobs that are that exists. We still need people programming, these machines, we still people need people maintaining these machines, it's just different jobs. Right. So so I think that we all have to have that culture of being flexible to retool as industry changes, and learning new things. And that's another thing that that's important of being having that mindset of lifelong learning. Because you're always be at the cutting edge of whatever it is that you're doing. And when you have to retool, you have that capacity that that mental brain flexibility to just learn something new. And you'll be fine.

Felipe Engineer  1:02:30  
Right? We're always going to need people.

Fernanda Leite  1:02:31  
Yeah, absolutely. 

Felipe Engineer  1:02:32  
Well, Fernanda, thank you so much for your time, I've learned a ton. I'm gonna have to rewatch this myself just so I can further my studies here and see what's going on in the industry. I really appreciate you spending time with us and sharing some of the research you've done with University of Texas and even before so thank you so much. You get the last word before you say goodbye to everybody.

Fernanda Leite  1:02:51  
Well, it's my pleasure. I'm thank you for for reaching out and for having this conversation. I really enjoyed talking to you. And I'm very optimistic about what the future holds for our industry. And I know that this industry can inspire an entire generation of kids to go into STEM and to to go into construction engineering because they see that it's, it's really cool. And we're doing we're doing robotics to we're doing automation, we're doing 3d printing, look at all this cool stuff. We can also inspire the next generation leaders to really want to to improve our industry as well. Thanks for having me.

Felipe Engineer  1:03:27  
Very special thanks to my guest. I'm Felipe Engineer Manriquez. The EBFC Show is created by Felipe and produced by a passion to build easier and better. Thanks for listening. Stay safe, everybody. Let's go build!