Betsy Cannon is a member of technical staff at Open Athena’s New York office, where she leads our materials science projects, RHOAR-Net (in collaboration with the Rosen Research Group at Princeton University) and MarinMat, which aim to further the discovery of new materials. In this Q&A, she explains what drew her to this work, her cookie-related introduction to LLMs, and what a project on electron density can learn from one about ocean modeling.
What’s one of your earliest experiences with AI technology?
Betsy Cannon: That’s a hard question, mostly because what we consider “AI” has changed. If I think back to the more recent LLM history, I remember testing ChatGPT when it was first released and trying to understand what it could do or not do. I like baking, and I had recently made a recipe that I make once a year. Usually the cookies turn out nice and crisp, but this time, they had spread. So I was like, “Let me just put this into the model.” I had a couple of instincts myself, and I wanted to see what it came up with. Overall, it agreed with my instincts: “Maybe I should have refrigerated the dough beforehand; maybe I could add a bit more flour to thicken the cookies up.” But it also came up with one other thing that made me think, “Oh, I can try that.” It’s interesting to see how the technology developed from confirming what I already knew, until now, when I can put way more into it in terms of code and letting it go much further.
What drew you to work at Open Athena?
Back before I went into tech, I majored in physics in college, because I always liked puzzles and understanding how things worked and trying to go to another level of deeper understanding. With physics, I really liked the modeling side and working with data, which is how—a couple years after graduation—I found my way to working in machine learning and building models for startups in the tech industry. So when I heard about Open Athena a year and a half ago, it felt like a great way to return to what I liked about physics: understanding what is out there, how our world functions, and trying to learn more and more about the universe. While not having to be the science expert myself, I got to learn from actual experts and share the expertise I built in industry. It felt like a fascinating problem space where I could contribute and also learn a lot.
Which Open Athena project do you primarily work on?
I work on RHOAR-Net, which is our materials science project, a collaboration with the Rosen Research Group at Princeton University. With that, we’re trying to accelerate the discovery of new materials by understanding their electronic properties.
Did you have any experience in this domain before you began working on the project?
I hadn’t worked in this particular problem space; I didn’t even know some of the basic terms. The first couple of meetings were me just wrapping my head around the project. But last week, I was at a conference where everyone there worked in AI for science. A few of them were also materials scientists, and I felt that I knew enough to hold my own. Not enough that I could push the envelope and be like, “Oh, this is the direction materials science could go,” but I could communicate what I’m building, what I’m working on, and have enough understanding that when people asked questions, I was like, “Okay, I see what you’re getting at, I understand what you’re trying to compare it to.”
What sort of things do you work on outside of a specific science project?
I’m an engineer. We like looking at problems that arise in different spaces and asking what can be generalized—what commonalities there are. So one thing I’m working on right now is looking across some of our projects and seeing, can we build on the same foundations? And does it even make sense to do that in the days of agentic AI, versus saying we should build separately and solve the same problem twice because AI makes the process fast enough? For example, I work with large amounts of three-dimensional data. So do the folks who work on our ocean models. I work on seeing if we can do things in common, what we can learn from each other’s projects, and starting to work more together, rather than simply having a bunch of piecemeal projects.
What do you like best about working here?
Great question. I think a bit is working on problems I didn’t know about beforehand. I have no background in materials science. I was a physics major, so I knew something about electrons, but very little. It’s also been a few years! I liked getting to learn about density functional theory, or DFT. It’s a key method for modeling electronic structure, which I had no idea about until late last year. Finally, I was like, “okay, how do we actually understand new materials? How do we know what we expect properties to be before we formulate a new one?” Understanding how this research is done has been fascinating. It’s like getting a window into new fields.
What does a typical work day look like for you?
It varies from day to day. I feel like I have a couple days of the week where I get to do more focused work. I might make sure the data is in the right place, try a new training run, or try to optimize something to measure a new benchmark. So a couple of days are focused on what my agents are running and making sure they’re moving forward and getting pull requests out the door.
And then usually I have a couple of days that are more meeting heavy. Those may be for internal planning—going back to how we work together, which things we have in common, learning more about the other projects—so I’m not only in my silo of materials, but also understanding what others are building that I should learn from, or learning from them about tools I could use. I also have a meeting with the lab once a week to understand where the research is going, what questions we should be asking, and how our progress compares to new papers that I haven’t read yet.
Do you have any favorite AI tricks or tips?
This isn’t a super advanced trick, but I will say I enjoy how much of pull request review can just be done by a bot. I have it already in GitHub, it reviews the pull request, and I can just have it go back and forth a few times. It’s nice—I can come in at the end and be like, “oh, it’s already in good shape,” instead of having to do quite as much back-and-forth along the way.
Do you have a favorite piece of science trivia?
I’m fascinated by biodiversity hotspots. For instance, Ecuador has over one and a half times as many bird species as the US, in spite of being roughly the size of Colorado.
Can you tell me about a hobby or activity you enjoy outside of work hours?
I love pottery. I’ve been doing it for over a decade at this point. It does unfortunately result in me having more ceramics than I need, so I appreciate friends who take those pieces away from me. I love being able to create new things, come up with new shapes, and push forward and learn more. And it lets me make something with my hands, something concrete.
Do you have a signature response emoji, or one you use most often?
I think most often, it’s the thumbs up, but I feel like that’s true for most of us. The one that I think differentiates me more is the waving fox that pops up. I think I probably use that one more than anyone. I also have a “yay llama” one with the arms going back and forth. So I’d say those are the two that I use disproportionately.
Cite this post
@misc{bushwick2026_meet_our_team_betsy_cannon,
author = {Bushwick, Sophie},
title = {Meet our team: Betsy Cannon},
year = {2026},
month = {jul},
howpublished = {\url{https://oa-www-demo.pages.dev/blog/meet-our-team-betsy-cannon/}},
note = {Open Athena Blog}
}