Spent a decade in tech, cut my teeth by learning JavaScript well when no one wanted to. Early in my career worked with cofounders of heroku on some prototypes, and that got me interested in end user computing and tool-making. After that moved to the Bay, did a stint at Apple working on internal tools, then moved to Amsterdam and worked at a smol cybersecurity startup. Got acquired by IBM, helped with the transition, then went back to startup-land, where after some pivots now we are working on an api development tool! Which I demoed at a tinkerers event over the summer. Been experimenting with LLMs and image generation since late 2022 which has kind of reignited my interest in technology. I feel like we’re rapidly unlocking new possibilities every few months. It’s awesome.

I’m very interested in the mechanics of chunking up a codebase in order to give an LLM the right context for a task. Also interested in people who have built assistants for specific stacks or infrastructures. This is to solve what I see as a limitation of current Code IDE Copilots — sometimes they don’t really understand (e.g.) that you might be writing code for a specific serverless environment. Finally, I’d also like to learn how to write robust evals for ai systems.

Right now I’m tinkering on how to quickly bootstrap Hono APIs on cloudflare workers, given a user’s description of what they want to build