Give Your AI Agent a Job Description
AI agents work best when treated like new hires: clear scope, limited access, and someone accountable. A worked example from an e-commerce platform.
3 min read
Secrets Sprawl: The Quiet Security Debt in Your Stack
API keys and tokens multiply faster than anyone tracks them. Here is why secrets sprawl is the debt nobody budgets for, and how to start paying it down.
2 min read
The Quiet Return of the Batch Job
We spent a decade making everything real-time. Then we bolted AI onto it and discovered that some work is happier waiting its turn.
2 min read
Your Agent Doesn't Have a Memory Problem. It Has a Filing Problem.
Every team that hits the wall on a long-running agent reaches for the same lever: a bigger context window. It is almost always the wrong lever. What is actually broken is the decision about what gets written down, what gets thrown away, and who decides.
10 min read
The Eval Gap: Nobody Can Tell You If It's Working
Most teams shipping AI features cannot answer the simplest question about them — is this actually good? Evaluation is not a testing chore. It is the definition of correct, and writing it down is the hardest and most valuable work in the project.
8 min read
Who Is Your AI Agent Logging In As?
Most organisations have given their AI agents real power and borrowed credentials. The identity question sounds like plumbing, but it decides whether an agent is auditable, revocable, and safe to expand.
9 min read
Semantic Caching: The Cheapest AI Answer Is the One You Don't Regenerate
Traditional caches only recognise a question they’ve seen letter for letter. Semantic caching recognises meaning — and for the thousandth shopper asking ‘does this run small?’ in slightly different words, that difference is most of your bill.
3 min read
The AI Gateway: Air Traffic Control for Your Models
As teams wire up a dozen different models, a new piece of plumbing quietly becomes essential — the AI gateway that routes, governs, and pays the bills. Here’s the architecture, with an airline booking example.
5 min read
Context Engineering Is the Skill Prompt Engineering Was Pretending to Be
The clever one-liner prompt is fading. The real craft now is deciding what an AI model gets to see before it answers.
3 min read
Small Models, Big Shift: Why Edge AI Is the Quiet Story of 2026
The headlines chase ever-larger frontier models, but the more interesting move in 2026 is downward — small, capable models running on the device in your hand.
2 min read
Building AI Features That Earn Their Place in Production
Most AI feature experiments share the same fate: impressive demo, uncomfortable production metrics, quiet retirement. Here’s the design thinking that changes the outcome.
2 min read
Zero-Downtime Migration: What the Case Studies Don't Tell You
Every cloud migration case study ends with success. None of them tell you about the Thursday at 2am when the database connection pool hit zero during a live flash sale.
2 min read
Your Digital Transformation Failed in Month One. You Just Don't Know It Yet.
After two decades of inheriting other people’s burning platforms, I can tell you exactly when enterprise transformations die — and it’s not when the CTO finally admits it.
2 min read