I build reliable, measurable, economical AI systems.
AI systems engineer focused on reliable, economical LLM products. I currently lead company-scale AI infrastructure and training work at SAP, after building ML platforms at Noom and distributed inference systems at Salesforce.
I write and build around reliable agents, evals, context systems, runtime architecture, and performance economics.
Current focus
- Reliable agent and LLM runtime systems
- Evaluation and post-deployment monitoring
- Context and retrieval systems that scale
- Performance, latency, and cost economics
Selected projects
All projectsLatticeDB
A single-file database that combines graph traversal, vector search, and full-text retrieval for AI systems.
Creator · Active
Nori
A distributed key-value store built to explore the mechanics of consistency, replication, and operability.
Creator · Research project
Quill
A protobuf-first RPC framework focused on transport pragmatism, developer ergonomics, and real failure semantics.
Creator · Research project
Writing
Writing archiveNo essays are published yet. The first pieces will focus on the practical problems that make AI systems succeed or fail after the demo.
- The eval pyramid for production agents
- Context is a systems problem, not a prompt problem
- What teams misunderstand about AI cost versus latency
- Why enterprise AI pilots die after the demo
Contact
Email is the best path for serious conversations. You can also find the project work on GitHub.