Engineer focused on AI enablement, agent infrastructure, evaluation, and developer tooling.
I build practical systems around AI-assisted software development: orchestrating and verifying coding agents, grounding model outputs in evidence, integrating tools through MCP, and automating developer workflows with explicit safety and recovery boundaries.
| Project | Focus |
|---|---|
| github-agent-router | Policy-driven routing, certification, and handoff loops for coding agents |
| aicommit-split | Deterministic semantic commit grouping with optional LLM planning |
| jev-ground | Evidence-grounded factuality verification for LLM and agent outputs |
| git-fleet | Safe, policy-driven automation for fleets of Git repositories |
| ticktick-mcp | MCP integration and task-workflow automation |
| activitywatch-calendar-export | Activity instrumentation and automation with AI-agent-readable outputs |
Open Connections PR #83 — storage abstraction and a mobile-safe vault-file persistence path for an embeddings-based Obsidian plugin.




