I'm Paul Bergeron, a senior software and systems engineer focused on applied AI. My work spans agent orchestration, evaluation and data infrastructure. I build across the application and the systems that support it, with an interest in making behavior observable and quality defensible.
Engineering responsibility
I lead the Agent + Tools team at SpaceXAI. My professional project details are confidential; the public reference systems and experiments on this site are separate work that you can inspect.
A scientific foundation
I earned a Ph.D. in Entomology at Washington State University. My research background includes ecology, integrated pest management, experimental design and molecular methods. Research taught me to examine whether a measurement actually answers the question being asked.
That habit carries into software: define the question, make the assumptions explicit, and check what the result supports. Explore my scientific work.
A teacher's approach
I've taught science and programming across secondary and higher education. I value explanations that help someone do something: test an idea, write a program, build an artifact or interpret a result. Explore teaching.
What you can inspect here
The software projects on this site are original public experiments. Each explains a mechanism, a consequential choice and a limitation. They are separate from confidential professional projects, which I don't discuss here.
The examples are intentionally bounded. A small training workflow can make the distinction between model suggestion and deterministic computation visible. A rubric can make evaluation assumptions inspectable. The useful question is what each example demonstrates—and what remains to be proven.
Get in touch
If you're building a system, evaluating applied AI, or looking for someone who can connect technical depth with clear explanation, send me a note. Tell me the problem, the constraints and the responsibility you need someone to own.