Journal
Field notes
Notes from
production.
What we learn shipping AI systems that have to run every day. No hype, just the work.
Strategy
6 min
Why most AI pilots stall in month four
The model is rarely the problem. Ownership, data contracts and a missing on-call rotation are.
Engineering
9 min
Shipping an LLM feature with a latency budget
How we keep p95 under 800 ms without giving up answer quality.
Field notes
7 min
What we learned from 80 production systems
Ten patterns that keep showing up in the systems that are still running.
Engineering
5 min
Evaluation before architecture
Write the test set first. The rest of the roadmap follows from it.
Next
Your pipeline
Bring us the
business problem.
We’ll tell you whether AI is the right move, and what it would take to make it work in production.