A self-hosted tool that scans a GitHub repo, runs an AI-powered gap interview, and generates a structured developer onboarding guide — so new hires reach their first commit without senior hand-holding.
A two-agent AI tutoring system: a Professor agent delivers structured lectures and quizzes with assessment gating, while a TA agent independently reviews lecture accuracy and quality using a second model — agents checking agents.
A multi-command CLI built on a shared tool-calling agent loop — each command (search, tailor, prep, practice, debrief) is its own agent with a scoped toolset, all running on one execution core. I built it to sharpen agentic engineering patterns like tool design and multi-step reasoning, using my own job search as the real workflow to automate.