Applied Scientist Lead (Solutions), Descartes Labs
Education: Ph.D. in Geography (University of Wisconsin-Madison), M.S. in Environmental Health Sciences (University of California-Berkeley), B.A. in Geography (University of California-Santa Barbara)
Describe your career path following your Ph.D. up to your current position. What are the range of tasks / responsibilities for which you’ve been responsible in the positions you’ve held?
I submitted my dissertation in December 2015, but I had already begun at a company called Descartes Labs, a startup that analyzes petabytes of satellite imagery. We have a team of software engineers that deals with ingesting the data, correcting it, and building a Python-based platform to access it. I’m on the Applied Science team, and we’re using the data to generate models or maps of various land cover types. Some of this is client driven; for example, a client might come to us and want to know where all of the corn in the United States is and what the yields will be at the end of the year.
We have a pretty wide variety of clients. We work with the U.S. government, and we have a particularly cool project with DARPA assessing food security in the Middle East and North Africa. We have also worked with Cargill – they’re interested in agriculture, obviously, and we also have other sorts of commodities clients. We’ve also done work with non-profits and have contacts at World Resources Institute and National Geographic.
The project that I’m currently working on involves using radar data to map rice across Asia, so it’s sort of a return to what I was doing in my Ph.D. I started at Descartes while I was still finishing my dissertation. I was looking at a lot of jobs as I was finishing, but I wasn’t really finding jobs that seemed like the right fit or where I wanted to be geographically. I Googled “private sector jobs with remote sensing” and this small company came up – the company was only 10 people at the time. I interviewed and really liked New Mexico, the people, and the team, so I decided I’d give it a shot and have really enjoyed it.
The job has evolved since I began working here. When I started it was only about a dozen people, and there wasn’t much structure on the team. Now we’ve grown to about 75 people, so we’ve had to get quite a bit of structure in place. I lead our solutions efforts; on the Applied Science team, we’ll have customers come to the Sales team and pitch an idea, and I help the Sales team assess whether or not the idea is technically feasible. We now have a proposal process in place, where we’ll do a more in-depth review of the project to see how we would approach it, roughly how much time it would take, how many people it would require, and then go from there. I lead the team that is doing all of the solutions work and then work with the Sales team to make sure we’re not committing to something we can’t do.
Can you talk a little more about the analysis you perform – you mentioned that it’s not just limited to the U.S., correct?
Yes, we have the entire historical Landsat archive, as well as global MODIS data and European Space Agency Sentinel data across the globe. When the company was just getting off the ground our initial project was trying to predict corn yields in the United States earlier in the season and with higher accuracy than the USDA, which we did, and that’s how we got our first couple of customers. From there, we’ve looked at a variety of applications including forestry, construction, and different types of agriculture across the entire globe.
The backbone of the company is the computing systems that we have. The team that started the company came from Los Alamos National Labs, and they were experts in high performance computing. They picked up on remote sensing quite easily, then built a system that can process data very efficiently so we can scale analyses really, really rapidly. I think I processed all of the data for Asia and ran that model in less than six hours for the entire continent. That’s really the powerhouse of the company.
How has your education/background in geography prepared you for this position?
I think the thing that jumps out most obviously is having a good grasp of procedural geographic knowledge. We work with spatial data and massive amounts of satellite imagery on a regular basis, so having a thorough understanding of these kinds of data is critical to what we’re doing on the job. For example, we have a project monitoring food security in the Middle East and North Africa. Ideally, we would just choose a random assortment of points and label all of them across the region to train and test our models. However, that’s not feasible given constraints on how you can move across the area. For instance, there may not be road networks, or you might be trespassing onto private land. Thinking about how to best use the data given, that it might be quite linear and not indicative of the entire region that you may want to be sampling, is very important.
I think understanding that these issues exist and being able to creatively brainstorm how to overcome them is pretty critical for what we do, as a lot of clients that we work with may not necessarily understand those concepts as thoroughly. They’ll often come to you with a vector dataset of points, and not understand why you can’t use those points right off the bat. In order to communicate that information to clients in a way that makes sense and is easy for them to understand, you need to have a pretty deep understanding of the datasets and what’s possible or isn’t possible given the constraints of the imagery that’s being collected.
How does your knowledge of the region under analysis – for example, the landscape characteristics or the cultural/economic geography of the area – inform your overall approach in how you use these technologies and how you analyze the environments?
It’s a huge component of what we do. The most obvious example I can think of is just the climate of a region. When I was working in Vietnam during my Ph.D., I was relying on Landsat data, but for six months out of the year, you don’t get a single scene where you can actually see the landscape because of the monsoons. At Descartes, it’s been much easier because we have all of the radar data from the European Space Agency on their Sentinel-1 satellite. Radar data can be difficult to work with, but in regions like the monsoonal tropics, it’s the only way you’re going to get data throughout the whole year. Having that knowledge of where you’re actually studying and understanding those climatic impacts is crucial.
Also, there are geo-political considerations. Looking at the tariffs that are now being imposed by the United States – is China going to start sourcing all of its soybeans from Brazil now? Keeping track of that and understanding how that might affect some of the models that we’re running is critical in having a deeper understanding of the dynamics that we’re trying to study and better understand. Because some of it you can’t just explain with the satellite imagery, you have to have a broader view of what’s going on globally with politics and trade.
What have you observed in your career in terms of positive impacts in the community or for your clients? How has geography enhanced the work of your organization?
In terms of the organization, I mentioned that the company was very small when I came on and most folks on the team had a background in astrophysics or computer engineering. It was amazing and pretty humbling what they were able to do without formal training in remote sensing or geography. However, being armed with some of the knowledge that I got during my Ph.D. has been important.
Understanding the terms that people are using in more detail has also been critical. For example, if we want to map urban areas, what exactly is an urban area? We think about urban areas being San Francisco or Washington, DC or New York, but globally those are quite different and varied. In Vietnam, for example, many parts of the cities are made of vegetation, so we aren’t necessarily picking up that signal in remote sensing data the same way we would looking at San Francisco. So, bringing those ideas of geography and what constitutes a certain land cover type, or this broader picture of what the data are saying has been useful for the team. I have since recruited some of my friends out of Madison so we now have more geographers on the team, which is great, as we’re able to convey geographic concepts to the broader group. I think it just arms us better to go out and talk to other experts in the field, and better communicate with our clients as well. For the organization, I think it has been very useful to have that knowledge on the team.
In terms of our community, Santa Fe is a small town and our company is a big fish in a small pond, so we’ve been able to give back in a variety of ways. We are out and involved in the community, and everybody thinks that what we are doing is mind-blowing, so it’s fun to present our work locally. We’ve done a few presentations around town and at schools, and it’s pretty neat to watch people who have no experience or exposure to satellite imagery see what sort of things can be done with it.
When you look back at your education trajectory, how did you discover geography? How did you realize that it connects with your aspirations and your goals in life, as a private citizen, but also as a professional?
I had a really, really amazing teacher for my intro to geography course at UC Santa Barbara. I didn’t know geography was something that I could major in when I went into undergrad; I took the course to fulfill a math requirement. The course was amazing. Carl Sundbeck, the lecturer for the course, showed different images of a road trip he took in the 1970s across the entire United States. He was describing the landscape and how different mountain ranges formed, or how a city such as Los Angeles grew up where it did because of the geography of the region and why that was beneficial. Drawing these connections just sort of blew me away, and suddenly the world made sense in a way I had never thought about. And it made me want to learn about geography, obviously, so then I ended up majoring in it.
I traveled for a bit between my Bachelor’s and Master’s programs, and it was amazing making those same connections that I was making during my initial intro to geography classes. It was really exciting and made me want to continue down that path. During my Master’s degree, I was doing a lot of data processing and epidemiology work, which is very statistical and very important, but it’s just not for me. I felt stuck at my desk and not really doing the work that I wanted to be doing. I went to Vietnam the summer between the two years of my Master’s program with a friend of mine from undergrad who was from there and we stayed with her family. I felt the same way as an undergrad when I was traveling. I wanted to get back to geography and do research that was in the field, which was a lot more dynamic than what I was working on for my Master’s thesis.
I then ended up going to Madison for my Ph.D. because the woman who taught my remote sensing classes at Santa Barbara had moved there. So I went to work with her and basically went with the excuse that I wanted to continue exploring, traveling, and getting to know other cultures. I proposed to do the work in Vietnam and got it funded by NASA. I was really lucky in that regard, as I got to sort of drive my own ship. And then coming to Descartes, it just sort of continued. I’m not necessarily out in the field collecting data and interviewing farmers, but it’s still every single day looking at satellite imagery and trying to better understand the world, so it’s exploration in a different form. It’s doing a lot of exploration via satellite imagery, not necessarily boots on the ground, but it’s still fulfilling that interest.
Do you remember having an “a-ha!” moment, that perhaps changed the way you think about issues and topics? Is there another example you might have?
It definitely occurred when I was at Santa Barbara. I realized that “Wow, you can earn a living doing something that you really love!” I don’t know what I envisioned I’d be doing after college when I was 18, or what I’d be doing as a career. There was some moment at Santa Barbara when it all clicked. I remember realizing there are jobs where you can continue on a path of scientific discovery and exploration, that you don’t just have to be crunching numbers and data or doing something that’s just a job. You can do something that you really feel passionately about. You know, it probably was that first geography class, and then the same lecturer taught the follow-up class the next semester. I think at that point I was pretty hooked. At that stage, I hadn’t even taken a remote sensing class and didn’t realize what career opportunities were out there. It was just sort of like: I love this; this guy can make a living doing this, so there must be room for other people. Those first couple of classes in college were pretty crucial to where I am now.