Engineering AI-Native
Ecosystems at the Edge
Founder of Stackbilt. I build interconnected AI developer tools, MCP servers, and deterministic inference engines—bootstrapped and deployed on Cloudflare's global edge.
What I Build
AI Infrastructure
Persistent cognitive systems that accumulate knowledge, route across models by task complexity, and operate autonomously — not wrappers, the layer underneath.
Developer Tools
Open source frameworks and automation that reduce the overhead of building and operating AI-native systems. Published, used in production, battle-tested.
Products & Platform
Live edge-native SaaS taking customers. AI image generation, architecture scaffolding, food data APIs, games — all bootstrapped, all on Cloudflare.
How I Think About This
Most "AI products" are thin wrappers around a single API call. I build the layer underneath — persistent cognitive infrastructure that accumulates knowledge, routes across models based on task complexity, ships its own code through governed pipelines, and operates autonomously with real safety constraints.
AEGIS runs 20+ scheduled tasks right now — ARGUS webhook intelligence with compound multi-agent consultation, agentic harness (compound casts with CTO/CISO circuit breaker), Nexus cross-repo contract awareness, dreaming cycle, cost monitoring, behavioral detectors — live, unattended, on edge infrastructure. It orchestrates headless Claude Code sessions that take a GitHub issue and deliver a reviewed, branch-per-task PR.
Everything runs on Cloudflare's edge. No origin servers. No containers. Sub-50ms globally. TypeScript end to end.