My Mac Setup Is Infrastructure as Code for One Person
A field note on using Homebrew, mise, and chezmoi to make a Mac reproducible without sanding off the human parts.
Read moreTheory is easy. Prod is hard.
Works in Prod
Building software that survives production.
Practical notes on AI engineering, architecture, testing, observability, developer experience, and the awkward bits that only show up after something ships.
Vibe-posted. Fact-checked. Mostly.
Start Here
Featured Posts
We Spent Five Weeks Making Docling Work. Then We Deleted It.
This is a post-mortem on five weeks of infrastructure work that ended with git rm and 1,452 lines deleted from the lockfile alone.
Read moreTDD Was Solving the Agent Problem Before Agents Existed
The first time I set an agent loose on a real codebase, it ran out of context before it had done anything useful. That's a clarifying experience.
Read moreBrowse by theme
Topics I Write About
AI Engineering
LLM systems, agent-readable code, and the limits that show up in production.
Architecture
Boundaries, diagrams, constraints, and decisions that outlive the sprint.
Testing
Feedback loops, architecture tests, TDD, and keeping drift visible.
Observability
Traces, metrics, latency, and making opaque systems explain themselves.
Developer Experience
Tooling and workflows that make setup, review, and delivery less theatrical.
Technical Leadership
Pragmatic delivery notes for teams trying to ship without lying to themselves.
Current focus
What I'm working on
Building
- TripForge
Learning
- Agentic systems
Community
- Software Crafters Manchester
Reading
- Designing Data-Intensive Applications
Latest notes
Recent Posts
My Mac Setup Is Infrastructure as Code for One Person
A field note on using Homebrew, mise, and chezmoi to make a Mac reproducible without sanding off the human parts.
Read moreThe Classic 'Works on My Machine' — Now With Neural Networks
There's a version of this story where we made a mistake and fixed it. That's true but incomplete. The fuller version is about what a structured evaluation process gets right, what it misses, and how the ground can shift under you even when you've done the work.
Read moreWe Spent Five Weeks Making Docling Work. Then We Deleted It.
This is a post-mortem on five weeks of infrastructure work that ended with git rm and 1,452 lines deleted from the lockfile alone.
Read moreThe Art of the Architecture Diagram Is Knowing What to Leave Out
There's a particular type of diagram that engineers love producing and nobody can read. It has boxes for every service, arrows for every dependency, labels on each arrow explaining the protocol, and a legend in the corner that requires its own legend. It is, technically speaking, accurate. It is also completely useless.
Read moreYour Codebase Has Rules. Does CI Know That?
There's a particular kind of meeting that happens on mixed teams. Someone's opened a pull request, and two engineers are staring at the same diff with completely different facial expressions. One is confused. The other is quietly furious. Neither is wrong, exactly — they just have entirely different mental models of what the codebase is supposed to look like.
Read moreYou Can't Debug What Bedrock Swallowed
There's a particular kind of hell reserved for debugging LLM-backed systems that nobody bothered to instrument. You've got a request that took twelve seconds and you don't know if the slow part was your retrieval pipeline, the prompt construction, the Bedrock call itself, or the post-processing that turned the model's output into something you'd actually show a user. You have logs. You have vibes. You have, essentially, nothing.
Read moreTDD Was Solving the Agent Problem Before Agents Existed
The first time I set an agent loose on a real codebase, it ran out of context before it had done anything useful. That's a clarifying experience.
Read moreThe Blockers Don't Care That You're Using AI
I wrote a post back in 2021 about walking skeletons — the idea that before you go deep on features, you ship something thin and deployable end-to-end. Not because it's useful to users, but because it flushes out the real blockers while the cost of finding them is still low. Permissions. Pipelines. Infrastructure assumptions that looked fine on a whiteboard.
Read moreThe Metric Your Users Feel Before You Measure It
Working on a streaming chat product taught me something: the standard latency metrics don't really describe what users experience. They're not waiting for a page to load or an API to return a JSON blob. They're watching tokens appear — and what they feel before anything appears is the thing most teams aren't measuring.
Read moreStop Arguing With Your Terminal About Python Versions
Project setup should be boring. Not in a "this is beneath me" way — in a "this takes thirty seconds and I never think about it" way.
Read more