A desire to constantly learn more has fueled Edana Martin’s varied career in science. Growing up in Colombia, she became fascinated with physics and astronomy, seeking to understand how the world around her works. Martin studied nuclear physics and completed a PhD at Universidad Nacional de Colombia where she simulated physical interactions between different particles of matter. Along the way, Martin taught herself a wide range of programming and data science skills.

Now, Martin enjoys providing for others the support she found missing in her research career. Following postdoctoral work at GSI Helmholtz, an ion accelerator facility based in Germany, Martin moved to the US and switched her focus from conducting research to helping others carry out theirs.

A data science support role at Massachusetts General Hospital allowed her to take advantage of the programming skills she developed during her PhD, creating simulations and working in the high-performance computing environment. Eventually Martin, always looking to grow her skillset, became the manager of that team, learning to guide engineers in their work.

This background made her a natural fit for her current role as a principal product owner of DevNull on the Information Technology Services team at the Broad Institute of MIT and Harvard, where she supports scientists working on projects that require complex computational technology.

We spoke with Martin about her work managing computational projects and her advice for budding scientists in this #WhyIScience Q&A.

Tell us a bit more about your work at Broad.

I provide technical guidance, empowering our engineers to do their best work. My team has ownership of cloud orchestration and high-performance computing, which is combining computing resources across multiple machines. Our environment is smaller than other institutions, but we need to make sure that it's revamped.

We also support platforms like the Genomics Platform, making sure they have computational support for their workloads. We’re wrapping up a refresh of our hardware, which is the first time we are officially including GPUs — graphics processing units that do advanced computational mathematics — for all Broadies. People are very excited about that because they can jump in and use GPUs for AI work and machine learning.

How will researchers at Broad use these new tools?

We have about five groups that are already testing GPUs for different purposes such as image analysis. If you have an image that has millions of points, like a PET scan or a CAT scan, and you need to do some kind of machine learning analysis on it — finding regions or tumors, for example — you need GPUs. When you want to model how proteins interact with each other or with a virus, you can model that using GPUs. There's also a lot of work with GPUs in natural language processing, to get data from clinical notes or to make sure medical record numbers match. All these projects have machine learning models today.

What are you most proud of in your work?

The mission of the Broad is really important to me. I really like to work for a place where I feel connected with the mission, and that started when I was doing nuclear physics. My first job was working for a mine detection project building a device to search for landmines in Colombia, and it was very fulfilling. I could see why I was doing this work: it’s going to be so impactful, it's going to save people, it's going to help find these terrible devices that are everywhere in my country, as well as many other countries. Even though nuclear physics seems not to be the most altruistic thing at first, I am always looking for the best usage of the technology for society.

How did you develop your data science skills?

For my PhD, I programmed simulations of physical interactions with matter in C++, which generates a lot of data. During my postdoc, I generated four terabytes of data per experiment, and I had to sort through that. Learning to analyze it was just a matter of necessity because nobody actually teaches you how to code or how to analyze your data. You just figure it out. Those are the skills that really help me today because you need to make sure you are creative and approach the problem in the best way that you can.

What are some takeaways from your time as a researcher that you carry with you in your role today?

Having the opportunity to work in a very diverse culture was really important. During my PhD I was collaborating with people from 20 different countries at the same time. Everybody has their own way to work, their own way to communicate. You need to be able to really understand each other, not only in language, but in the way that people bring things up. It's really amazing when you have the opportunity to learn from different cultures. Not everybody has that opportunity, so I was very grateful for that.

What do you find most rewarding about your work? 

Coming from a science background, I always struggled with not having support. When I was doing my research, I didn't have anybody to ask. I didn't have anybody to help me store my data or teach me a better way to run code or how to make testing or debugging code easier. Having the opportunity to be that kind of support for other scientists is the most rewarding thing.

Do you have advice for someone starting out in a science career? 

I love science, and having the opportunity to learn what you're passionate about is really invaluable. But you always have to think about what is next and how your career is going to play out. When you focus your exploration of science so narrowly, you can close a lot of career paths. So I really encourage everybody to love your science, but never close your path because you never know where you will end up.

This conversation was edited for length and clarity.