Building tools for the boring half of AI.
Audit trails, evidence chains, traceability — the regulated-industry workflows that gate real AI adoption in automotive, aerospace, medical and industrial.
Twenty-three years of safety. Two years of AI. One stack.
I authored the architecture DSL that now generates ASIL-D braking and automated-driving software shipping in 10M+ vehicles. Today I'm building its AI-native successor in the open — so frontier models can ship inside the audit workflows that gate adoption in automotive, aerospace, and medical.
Open-source primitives for regulated AI.
Sylang
AI-native MBSE language for ISO 26262 / DO-178 / IEC 62304. Open-source successor to the DSL behind 10M+ vehicles.
Akela
Open-source agent orchestration with skill routing and an outbound-only adapter — agents stay behind the firewall.
Hermes Studio
Browser IDE for agents with kernel-isolated workspaces — true process isolation, not prompt-enforced. MIT-licensed.
hermes-adapter
Sidecar that exposes hermes-agent over A2A JSON-RPC and a workspace HTTP API. The connective tissue Akela and Hermes Studio use to drive stock upstream — zero fork maintenance.
JotX
Personal knowledge hub that syncs to your own Google Drive or OneDrive. No JotX account, no JotX servers — notes stay in storage you already own.
23 years building safety-critical systems.
Director, Engineering Systems
BorgWarner · 125+ products, 16,000+ users globally
Director, Systems & Safety Engineering
Luminar Technologies · Built team 3 → 25+ in <12 months
Head of Platforms & Systems
Continental Automotive · Authored the DSL behind 10M+ vehicles
Software Engineer
Infosys Technologies · India & Germany
Let's talk.
Open to conversations about industry AI for regulated workflows, MBSE, agentic platforms, or anything in this neighborhood.