Work / Experiments
Experiments
Exploring AI tools and infrastructure that empower developers.
Experiments collects focused prototypes around AI-assisted programming, evaluation, and infrastructure.
Ongoing
Research / Builder
Language models, evals, agents, developer tools
Overview
Experiments collects focused prototypes around AI-assisted programming, evaluation, and infrastructure.
Why it matters
Useful developer tools need tight feedback loops. These experiments test what is practical before it becomes product-shaped.
System
01
Prototype loops
Builds small tools to test concrete developer workflows.
02
Evaluation harnesses
Measures behavior with repeatable tasks and checks.
03
Infrastructure notes
Documents patterns that survive real implementation pressure.