Developing PostPython, a Python-to-C ahead-of-time compiler, with a focus on code generation, numeric-type correctness, and useful unsupported-feature diagnostics.
- Built accuracy, edge-case, and mutation-testing infrastructure for 27 statistical kernels; cross-validated Python, native shared-library, and NumPy ufunc results against SciPy and NumPy.
- Applied agent orchestration with Claude Code and OpenAI Codex in development workflows; wrote technical documentation and release notes and coordinated interns producing technical videos.
Leading development of Python pipelines that process terabytes of calcium-imaging video and EEG/electrophysiology data for a nine-person neuroscience research team.
- Re-architected the analysis package into extensible object-oriented components and introduced GitHub Actions CI and comprehensive end-to-end testing.
Mentored more than 100 students in Python, recursion, object-oriented programming, and software testing through labs, office hours, and individual code reviews.
- Reviewed more than 200 programming projects and provided actionable feedback on correctness, algorithmic efficiency, code quality, and style.