Start here
I build AI-native, multi-tenant data platforms and agent systems in production — from open-source Rust runtime substrates to live commercial deployments. This page is the shortest route to my best work and to me.
What I build
AI Lead at Stimulus (governed agentic procurement) and Data Architect at TEAM International; previously Co-founder & CTO at Wedi Pay (agentic cross-border payments, 2024–2026). On the substrate side: the Rust Agent OS stack — Life, Symphony, the control metalayer, bstack + broomva/skills, and the RCS paper series.
Why it matters
Most failures are not model failures. They are harness failures. I focus on the systems-level primitives that make agents controllable, observable, and useful under real constraints — with audit-grade governance as a default property, not an afterthought.
The full profile →
Concurrent leadership, anonymized client engagements, the OSS substrate, and downloadable CVs in three lengths.
Best projects
All projectsLife
ActiveThe Agent OS monorepo — 8 Rust workspaces, 50+ crates, 136K lines, 1045 tests. Arcan, Lago, aiOS, Autonomic, Haima, Praxis, Vigil, and Spaces unified under one roof.
March 15, 2026
Autoany
ActiveEvaluator-Governed Recursive Improvement (EGRI) — turn ambiguous goals into safe, measurable, rollback-capable optimization loops over any mutable artifact.
March 12, 2026
Symphony
ActiveRust-native orchestration runtime for composable agent workflows with deterministic scheduling and fault isolation.
March 10, 2026
Best writing
The Benchmark Ran Out Before the Loop Did
We pre-registered a self-improvement experiment, and it was over in one generation — the model aced every task and the selection gate became mathematically unsatisfiable. Two weeks later another lab hit the same wall. The instrument didn't fail. The instrument class did.
July 21, 2026
The Folklore Arrived Before the Physics
In six months the field converged on a belief — the harness, not the model, is what matters — and backed it with striking empirics. What it doesn't have is a single theorem about when the loop it now worships converges, oscillates, or collapses. A map of the gap, and the four results that would fill it.
July 21, 2026
Forking Houston: 89 commits, six agents, one hack camp
We took Houston — an MIT-licensed agent runtime — and grafted a new brain onto it. In a single hack camp the fork went 89 commits ahead of upstream: four net-new Rust crates, a full Linear integration, a feature-flag platform, a push pipeline. The features are the easy part to show. The real story is the method that produced them — autonomous agents working in parallel git worktrees under a governance harness that made the speed safe instead of reckless.
June 6, 2026