Building coding agents, useful AI tools, and systems people can trust.
One task. Multiple models. Coordinated agents.
I'm building Coderouter, an autonomous coding agent that routes work across AI models and coordinates agents from planning through implementation and validation.
- Route intelligently — Match each task to a model based on capability and cost.
- Work in parallel — Coordinate agents in isolated Git worktrees and integrate their changes.
- Check the results — Run validation and keep model assignments, costs, and outcomes inspectable.
| Project | My work & why it matters |
|---|---|
| VerifyWise | Built LLM evaluation tools that help teams assess quality, hallucinations, bias, and toxicity—and build more trustworthy AI systems. |
| AssistAI | Built enterprise retrieval pipelines that help automotive employees find and use their company's information. |
| ForesightAI | Built forecasting and AI reporting systems for distribution networks spanning 1,400+ endpoints, helping teams understand performance and allocate resources. |
Have something interesting to build? Let's talk.