blog10 posts3 categories11 active topicsfiltered: ai

Notes from the field.

Working notes on the craft — engineering at scale, the AI-era developer experience, frontend that most teams leave on the table, and whatever else has my attention this week.

Building fedbench: an LLM eval harness for grounded Q&A
Most LLM agents that read documents and answer questions fail in three specific ways. fedbench measures all three. Here's what it does, why I built it, and what I learned shipping it.
ai
evals
rag
open-source
May 9, 202614 min
Building fieldops-mcp: designing the tool surface an agent actually uses
I built a small MCP server that turns a fictional dispatcher's workflow into agent tools. The interesting part wasn't the code. It was the design choices that decide what the agent can do at all. Here's what I built and what shipping it taught me.
ai
agents
mcp
tool-design
open-source
May 9, 20267 min

Building fedbench: an LLM eval harness for grounded Q&A

Building fieldops-mcp: designing the tool surface an agent actually uses