I build Arbor, graph-native code intelligence for AI agents and engineering teams.
Most AI coding tools guess from text. Arbor maps the codebase as a real graph, then gives agents deterministic context: call paths, blast radius, Git diffs, and MCP tools.
build deeply -> measure honestly -> ship relentlessly
- Cofounder building Arbor as the core bet.
- Focused on code intelligence, agent observability, and local-first systems.
- Preference: deterministic engines, fast feedback loops, interfaces that stay sharp.
| Project | Signal |
|---|---|
| Arbor | Graph-native code intelligence for AI coding workflows. |
| LoRA-JIT | Local adapter routing and paging telemetry for code generation. |
| Agent Lens 2.0 | Visual debugging for AI agent workflows. |
| J.A.R.V.I.S. | Windows-first desktop AI assistant. |
| FinShield | Financial document forensics and graph investigation. |
| witr | Process, port, and service causality diagnostics. |
Arbor / Repositories / Email / X





