What I Do
Engineering: full-stack from specs to UX — auth, databases, CI/CD, observability, clean APIs.
Research: bilevel optimization, implicit differentiation, A100 training runs, reproducible experiments, publication-grade code.
Both inform each other. I write systems that survive contact with real data. I write experiments that survive contact with real constraints.
Operating style
Deep dives across software, models, and systems.
I like work where a topic is worth digging into properly: the model, the data, the interface, the infrastructure, and the trade-offs that connect them. Surface-level understanding does not interest me much; I want to know why a system works, where it fails, and how to make it usable.
The intersection of fields is what keeps me engaged. Computer science gives me leverage, AI/ML gives me a way to model messy behavior, and my minors in physics and statistics push me toward stronger reasoning about uncertainty, structure, and evidence.
Why Engineering and Deep Learning, Together
CS gives leverage. Deep learning gives expressive power. Engineering makes things work. Research makes things correct.
I'm most engaged when both are in tension: when a model needs to be theoretically sound and practically deployable. When code needs to be reproducible and production-ready.
Infrastructure and evaluation matter as much as the architecture. Strong baselines, reproducible runs, measurable improvements.