AI Engineering

Every AI Engineering article we've published — field notes on building and shipping production AI systems, written by the engineers who build them.

6 articles

Harrison Itotia10 min read

Why Go is the perfect language for building agents

The model is the easy part. An agent is a long-lived concurrent server that fans out to unreliable I/O, streams while it works, and must never forget whose data it is touching — which is Go's home turf, not Python's.

Sep 2, 2026 Read article

Patrick Waweru10 min read

What an AI agent actually costs — and how to cap it

Token spend is the line item nobody scopes until the bill arrives. How we meter every agent server-side, budget it per agent, and stop runaway spend with a daily kill-switch.

Aug 15, 2026 Read article

Jane Njeri9 min read

Design is a model capability now - and why that changes agent UX

Frontier models have real design taste. Constrain it to your design tokens and components and the model becomes a fast implementer of your system - on-brand by construction.

Jun 24, 2026 Read article

George Onyango10 min read

Swapping the brain - building agents that aren't married to one model

The reasoning model is the most replaceable part of an agent. How we keep it swappable: a thin model boundary, an eval harness that measures agent work, and guardrails that don't move.

Jun 10, 2026 Read article

Joy Mwende11 min read

Generative UI - letting Gemini 3 design the screen

Generative UI has split into two techniques — a bounded component registry, and interfaces coded from scratch. When to reach for which, and the registry pattern end to end in Go and Vue.

May 13, 2026 Read article

Joe W10 min read

Agent search - wiring retrieval into an LLM agent

Your agent isn't bad at search - your search tool is. Why "search" is three operations collapsed into one, and how three legible gRPC tools let the agent compose them without touching the prompt.

Apr 23, 2026 Read article