Ugur Basak
Speaking
Talk ·
AI Observability Platform Engineering FinOps

Stop Treating AI Like Magic Infrastructure

Observability & SRE Summit (co-located with the Platform Engineering & Developer Experience Summit) · London

AI is a workload, not magic infrastructure. Why the observability, FinOps, and governance disciplines we already trust are exactly what AI and agentic systems need.


Most teams treat AI as a new universe with its own rules. It is not. AI is a new workload, and the disciplines that made the rest of the software lifecycle reliable, observability, FinOps, and governance, are the same ones it needs.

This talk makes the case for putting LLMs and agents on the same Golden Paths your services already run on, and looks at the operational blind spots that open up when teams skip those disciplines in the rush to ship.

What the talk covers

Why AI gets treated as magic infrastructure, and the operational blind spots that creates at scale.

Putting AI on the Golden Paths you already have, instead of building a parallel, ungoverned stack beside them.

The three disciplines that make it real: observability so you know what it is doing, FinOps so you know what it costs, and governance so you know who owns the decisions.

Also at the summits

Across the two co-located events in London, I am also part of a roundtable on whether a platform is genuinely ready for AI, or just exposing the friction that was already there.

Who this is for

Engineering and platform leaders putting AI and agentic systems into production, and the teams responsible for keeping them observable, affordable, and accountable.

Invite me to speak

I speak on Platform Engineering, Developer Experience, Engineering Leadership, and Organisational Transformation.

Invite me to speak