Posts
The repetition trap in grammar-constrained decoding
A faster GPU turned 'too slow' into 'broken': why grammar-constrained JSON extraction sends local vision models into repetition loops, and what helps. read more →
Pydantic AI structured outputs and evals on Bedrock
From schema-guaranteed structured outputs on Bedrock to Pydantic Evals: how I get LLM output I can actually gate a merge on. read more →
Building and securing MCP servers with FastMCP
Build a production MCP server with FastMCP 3: auth, per-group tool access, audit logging, agents behind tools, S3 signed-URL file delivery, and offline testing. read more →
Skills or MCP servers: when you need a server
Skills run in your own context with your own credentials. The moment data is shared and access has to be governed, you need an MCP server in the middle. read more →
Anchoring specs to code with ast-grep
A lightweight convention for binding spec sections to code structurally, so agents can find the right spec before a change and patch it after. read more →