Writing
Notes from the work
Long-form articles on the decisions behind the systems I build, plus the stories behind them. I post when I have something worth saying.
- Where the LLM goes in regulated reinsurance: a reference architecture for the API boundary and the audit trailEvery LLM tutorial assumes you can send the data to the API. In regulated reinsurance that assumption breaks before the first prompt — most of the architecture lives in the boundary, not the model.
- Vector DB vs. structured store vs. live API: where each piece of a serverless RAG actually belongsMost serverless RAG architectures fail not at the model but at the data layer — by trying to make the vector DB carry weight it was never designed for. The interesting decision is which piece of context belongs where, and the question that decides it is freshness.
- Designing the Modernization Factory: an 8-stage agentic pipeline for COBOL→AWS, and where each Claude model earns its placeTurning mainframe COBOL into AWS-native code with an LLM is easy to demo and hard to make repeatable. The architecture that makes it a factory rather than a party trick is in the pipeline shape, the per-stage model routing, and the one checkpoint that sits before the irreversible work.
- Grounding a career site in a vector DB: a $10/month chatbot that has to tell the truthBuilding /chat on my own site forced three decisions I usually get to defer on someone else's product: where the cost ceiling lives, what counts as a citation, and how to keep an embedding pipeline from breaking a deploy.
- Mainframe modernization without a re-platform freeze: a pattern from five years of rating-engine workEvery mainframe modernization plan dies on the same rock: the business can't tolerate a write freeze, and the engineers can't tolerate live dual-writes against a system they don't fully understand. The way out isn't faster cutover. It's separating the rules from the engine.
- Three minutes to five seconds: where the cache belongs in a serverless reinsurance appThe query was correct. The database was sized. The bug was that the cost lived in the wrong layer — and the fix was a cache that became load-bearing for correctness, not just speed.
New posts when there is something worth saying. Subscribe by RSS