I build production-grade systems: high-performance inference, distributed orchestration, secure execution, and economic engines that make compute sustainable.
- High-performance / low-latency design
- Containerization (Docker/Podman/Orbstack) + orchestration (Kubernetes)
- Observability-first architectures
- DevSecOps constraints & sandboxing
- Inference pipelines (ONNX/ORT, vLLM, Candle)
- Agent orchestration & tool ecosystems
- Retrieval + memory systems (pragmatic, not academic)
- CAC/LTGP + capital-efficiency modeling
- Compute tier governance + subsidy logic
- “Automation dividend” modeling and deployment economics
- Agency — Rust microservices constellation for autonomous reasoning + memory + voice, designed for concurrency + isolation.
- Compute Commons — framework for decentralized GPU coordination + economic governance.
- OpenAPI MCP Server — universal OpenAPI → MCP tool generation for high-speed agent toolchains.
- infra-as-an-organism — infrastructure treated as a self-aware system: feedback loops, governance, adaptation.
- ltgp_cac_calculator — Rust CLI economics engine for go-to-market & capital allocation.
- Personal AI Infrastructure — a pragmatic, deployable local inference stack that balances performance, cost, and control.
Languages: Rust, TypeScript, Python, C++
AI/ML: ONNX, ONNX Runtime, Candle, vLLM, retrieval pipelines
Infra: Docker, Podman, Kubernetes, edge/VPS clusters, GitHub Actions
Tooling: CLI-first design, OpenAPI automation, agent orchestration
- System-level leverage over feature velocity
- Performance as a first-class constraint
- Economic resilience as an architectural requirement
- Open ecosystems over closed silos
📍 Los Angeles, CA
Building systems that scale technically — and economically.