Field notes on the engineering side of AI and its delivery promises.
How teams adopt agentic coding without losing control and how you decide who owns which rules (all with a lot of Team Topologies, Domain Driven Design, Wardley Maps and more).
I’m Thiago de Faria and this is Proof in the Pipeline: AI governance shown working where the code ships, not promised in a pdf or powerpoint talking about “AI policies”.
20+ years ago I was a statistician in Brazil. Then a data engineer, then a CTO, a Solutions Architecture leader at AWS and now I run TilinTheCloud as an independent consultant on Data & AI Architecture who still writes the code (or should I say I plan-design-validate code at this time and age?).
Most companies now have an AI policy, yet very few can show it working. Their teams ship code with agents in days, but review, testing and approvals still run at human speed… and nobody is sure who is allowed to decide what.
So why is this so hard to fix? Because it’s a three-body problem. Architecture, teams and AI capability keep pulling on each other… move one and the other two move with it (hello, Conway). You can’t fix them one at a time with a policy on top. You steer all three together.
This is what you can expect, weekly:
Field notes from real delivery work… where the guardrails go in the pipeline and what changes when agents write the code.
Research (named and linked) to read against what I see in the terminal.
Team design under AI… cognitive load, decision rights and why governance fails when only one of the three bodies moves.
The weekly posts are free and stay free. If you want more, there will be a monthly paid post with more code and research updates behind the essays. Founding members (12 seats) also get a quarterly call with me and priority on the Delivery Lab (oh yeah, pledging costs nothing until paid launches).
I’ve done this work inside and alongside AWS, Nike, ASML, KLM, Rabobank, Booking and more. My July 2026 Executive MSc research at Antwerp Management School tackled how GenAI is changing the Software Architecture practitioners and this new distributed cognition paradigm.
I can ensure you no futurism. And never arrows and boxes without an actual test, code or recommendation.
If your agents already ship code, you’ll feel at home here… but I might throw some curveballs and some things for you to reason and consider.
And if you are reluctant to have a fully autonomous software pipeline, I also think you’ll feel at home here. But with arguments, science and practice we might be able to accelerate and get more outcomes as we speed the output.
Subscribe, interact, chat, learn and teach others (including me!).
Cheers,
Thiago de Faria
