Home / Insights
Ideas worth shipping.
Field notes from building AI and software in production — architecture, strategy and the hard-won lessons in between.
Beyond the chatbot: designing AI agents that actually do the work
Most enterprise AI projects stall at the demo. We unpack the architecture patterns — tool use, memory, evaluation loops — that separate a slick prototype from an agent you can trust in production.
Read articleThe data foundation your AI strategy is missing
Great models fail on messy data. Here's how we design pipelines that keep AI honest.
Read article IndustryRAG in regulated industries: a practical compliance playbook
Retrieval-augmented generation done right for healthcare and finance.
Read article Cloud & DevOpsShipping ML to production without 3 a.m. pages
The MLOps guardrails we put around every model we deploy.
Read article StrategyBuild vs. augment: choosing the right engagement model
A founder-friendly framework for how to actually staff an AI build.
Read article Generative AIFine-tuning vs. RAG vs. prompting: a decision guide
When each approach wins — and how to combine them for accuracy and cost.
Read article MobileNative or cross-platform? A 2026 decision framework
Swift/Kotlin vs. React Native and Flutter — how to choose for your product.
Read article SecurityDesigning AI systems that are secure by default
Guardrails, access control and audit trails for production AI.
Read articleThe occasional, useful email.
New essays on AI engineering and product — no spam, unsubscribe anytime.