Research Fellow and healthtech enthusiast!

An experiment in designing a refillable notebook through conversation, small prints, failed assumptions, and a much simpler MVP.

From a one-minute experiment to a six-minute product tutorial: what I learned about AI narration, browser automation, and making a video people can actually follow.

How we integrated LLMs into a conversational expense tracker — tool use, vision, voice, token-aware batching, rate limiting, and the engineering that makes it all work.

How I use a Lima VM for isolated Claude Code automation, and a wrapper repo pattern to version-control my Claude configs without touching the team's codebase.