Founding partner at Initial Capacity, a software consultancy of lead developers building applications at scale for startups and the world's largest companies.
I teach Foundations of Software Engineering at the University of Colorado Boulder. I hold a PhD in mathematics from CU, where my research focused on algebraic combinatorics.
Before starting Initial Capacity, I spent eight years at Pivotal Labs shipping production software, building engineering teams, and helping launch the Platform Acceleration Lab. I also started the Pivotal Labs Dublin office.
I write Go, Python, Kotlin, and TypeScript, among other languages. Most of my time is spent building production systems, with a current focus on AI applications, evaluation methodology, and search infrastructure.
I think a lot about how experienced engineers work effectively with AI coding agents, how to evaluate search strategies beyond embeddings, and how small teams of strong developers can outperform much larger ones with the right patterns and tools.
I write about software engineering at InfoQ:
- A Simpler Testing Pyramid: Getting the Most out of Your Tests
- The Unit in Unit Testing
- You Don't Need a CSS Framework
- Streaming HTML: Asynchronous DOM Updates without JavaScript
I speak at conferences and meetups about search strategies, information retrieval, and building software at scale. My current talk, "Embeddings aren't a search strategy," covers how keyword search, hybrid search, and filtering often outperform vector embeddings when paired with proper evaluation.
