Blog

Long-form posts on the things I'm building and the papers I'm reading.

1. Jun 6

   Three changes that made my coding agents reliable

   Engineering

   I spent weeks building a multi-agent harness for software delivery. The biggest surprise: as the models get better, the leverage moves out of the model and into the system around it. Three changes moved my reliability more than any prompt tweak.

   5 min read
2. May 20

   A drone that holds three minds in parallel — and only picks one when forced to

   Research

   Most autonomous systems force a single resolved output. Biology doesn't. CANavigator runs three navigation algorithms in parallel and only commits when the environment demands a response — and it beats the fastest single algorithm on both time and energy simultaneously.

   9 min read
3. May 15

   $65 a month to keep a portfolio demo alive

   Engineering

   How GCP Cloud Run's minimum instance billing cost me $65.75 in month one, what I changed, and what AWS had waiting on the other side.

   7 min read
4. May 14

   Two-stage retrieval, and why a 22M-param reranker beats a bigger embedder

   Engineering

   Building ProfessionalRAG: how I picked BGE-base + an MS-MARCO cross-encoder over a single dense retriever, and what the latency / cost / faithfulness tradeoffs actually look like in production.

   8 min read