Building at the source
Upstream
Now
Updated Jul 7, 2026
Building autonomous agent infrastructure: multi-agent coordination, design systems, and the tools that make AI engineering practical. Currently publishing Running a Swarm, a series on operating fleets of AI agents, and growing Recall, my agentic engineering vocabulary.
The Stream
When Failure Stops Looking Like Failure
Part one of this series made the case that the brief you hand a fleet agent is an interface: a typed, validated contract rather than a paragraph of hopes. You can check a brief automatically, before …
Source
The Brief Is an API
Running a swarm of AI agents, part 1: why the brief you hand each agent is an …
The River, and What It Cost
I built the research river and let it run unattended. The failures taught me …
Research Once, Render Many
Most AI work glues research to a single output and pays for the thinking again …
We Went Shopping for AI Research Tools. The Literature Told Us to Keep Our Wallet.
I sat down to upgrade my research toolkit and the literature told me to keep my …
Tests-Green Is Not the Truth
A green checkmark from the agent that did the work is not proof the work is …
The Library That Pays You Back
Most research is write-once, read-never. The fix sounds too obvious to bother …
One Agent, Many Questions
If one research agent is good, why not run ten at once? The answer is more …
I Set Out to Fine-Tune a Model and Ended Up Building a Research Wing
I set out to learn how to fine-tune models with AI. Little by little it became a …
Intent Engineering, Part 3: Grade the Grader
Can you trust a model to grade a model? I built the judge anyway, and it lied to …