👋 Hi there! I build ML apps for both the cloud and private clusters - hit me up if your team is deploying agentic workflows or NLP+vision tasks with data privacy features.
Check out my resumé and projects below ☕
Read everything before you sign, or vote
You may suffer from social media bubbles. Your AI suffer from basins of attraction.
So you want to have your own AI, and your own data kept private, and your own hardware utilized.
Eval tools for agents to eval its own eval? How to eval the eval?
Or how to use your home GPU when you’re at Starbucks with Docker, Traefik, and Tailscale
I used to be afraid of logits. I thought ML models’ outputs were labels or numbers - so what’s the deal?