Platform engineering for the AI era
Field notes on platform engineering for the AI era. DORA 2025 analysis, the three persona platform user, and the Foundations Framework in practice.
Cognitive Load diagnoses the problem. Cognitive Absorption names the fix. Four signals and why the distinction changes what platform teams are accountable for.
Most DORA implementations produce numbers that cannot survive a definitional question. Four failure modes explain why — and what to fix in each case.
How much cognitive load is your platform absorbing for your developers, and how do you know? Most engineering metric frameworks cannot answer that question. This post explains why, and what to instrument instead.
Most SLO programs fail because the targets are detached from real user experience. A practical guide to SLO design that engineering teams will own, not ignore.
Every hour engineers spend fighting deploy friction, waiting on platform tickets, or repeating slow onboarding is a real cost. A framework for making the number concrete.
An honest analysis for a VP Eng facing the build-the-team-or-bring-in-a-consultancy decision. Cover the 3-6 month critical window, failure modes of each approach, and what a good engagement exit looks like.
A structured decision framework covering total cost of ownership, team capacity requirements, vendor lock-in spectrum, what changes at 10 vs 50 vs 200 engineers, and the hybrid path.
Ad-hoc platform engineering is the industry default. This is an honest comparison of what structured methodology changes versus what ad-hoc produces over time — and who actually needs a framework.
Both frameworks are legitimate. The question is sequencing and context. DORA measures flow and stability. SPACE adds developer satisfaction, communication, and efficiency. Here is how to choose where to start.
Not a feature table. A decision framework for engineering leaders who have already accepted they need an IDP and now need to pick one without regretting it in 18 months.
Platforms in 2026 serve three user types, not one. What the AI Agent and Hybrid Collaborator personas need — and how platform design changes for each.
Growth-stage teams do not need Google-scale SRE. Three conditions say when it pays. Four capabilities in sequence address the reliability problems you have.
Signal Integrity: metrics that resist gaming and evolve with tooling. Three properties DORA alone cannot provide, and the adversarial review that builds them.
Long-form analysis on platform engineering, DORA, and AI readiness from Mat Caniglia. Sent when there is something worth reading.
Start here
A fifteen minute self-diagnostic that scores your platform across DORA metrics, deployment frequency, change failure rate, and cognitive load. No sales call required.
Want to read first? See the Foundations Framework