Platform Engineering Santiago (Chile)

Backstage gives you a beautiful developer portal. But a portal is not a platform.

Behind every internal developer platform sits a long list of decisions and integrations. CI/CD. Observability. Environment management. Deployments. Access control. Backstage gives you the front door and a plugin model to wire it all together. The wiring is where the real work lives.

Join Kelsey Hightower and the OpenChoreo team for an honest look at what it takes to build a complete internal developer platform. We start with the vision: what developers should get from a platform, and what platform teams have to build to deliver it. Then we get concrete. We map the Backstage plugin landscape, show how vanilla Backstage has evolved, and count the glue it takes to turn a portal into a working system. We compare a few common approaches to observability and CI/CD, and show where each one leaves you.

We end with the decision every platform team faces. Build all of this yourself, or adopt something that ships it as one system, like OpenChoreo. For most teams, that means fitting new pieces into a platform they already run, not starting over. We won't tell you which is right. We'll make sure you know what you're signing up for either way.

You'll leave with a clear map of the moving parts, a realistic view of the effort involved, and a framework for the build-versus-adopt call.

Speaker: Kelsey Hightower

Every postmortem asks what went wrong during the response. What often gets missed is the foundation that was already there before the page fired: the ownership, the runbook, the last deploy. We'll follow one incident from page to postmortem, then go back to the platform work that made it possible, including how to use AI to augment incident response.

We'll cover:

• What good looks like end-to-end: one incident from page to postmortem, investigated, remediated, and documented in one place
• The foundation underneath it: ownership and on-call in the service catalog, runbooks and dependencies attached, readiness tracked on scorecards
• Where AI actually augments the response, from catching responders up to drafting the postmortem, and the platform data it needs to do it

Speaker: Sam Rodman
Senior PM | Incident Management @ Datadog

Speaker: Jessie Wu

After two years of hype, many enterprises feel they’ve overspent on AI infrastructure without realizing the promised value.
In this session, Dan Ciruli, Kelsey Hightower, and Luca Galante cut through the buzzwords to show what enterprise AI infrastructure actually looks like in production.

In most cases, it is often less about training models from scratch and more about how platform engineering teams can fold AI inference and fine-tuning directly into their existing infrastructure.

Join the discussion to learn:

1. Why you don’t need infrastructure for training models from scratch, and how to start treating inference and fine-tuning just like serving standard web APIs.
2. Strategies for folding AI workloads and autonomous agents into your existing heterogeneous environments (VMs + containers + legacy databases) without rebuilding everything.
3. How to build an Agentic Development Platform (ADP) based on actionable reference architectures to deploy real-world AI applications.

Speaker: Kelsey Hightower
Retired Distinguished Engineer, Google Cloud. Author, Investor, and Technical Advisor

Speaker: Dan Ciruli
VP and General Manager, Cloud Native

Speaker: Luca Galante

ProsperOps surveyed 100+ platform engineers on managing AI cloud costs. Learn insights from our survey results. Together with our partner, Pay-i, we show you a practical framework for measuring AI ROI, avoiding bill shock, and building cost guardrails that scale with agentic workloads.

Every platform team is now tasked with funding and managing AI initiatives, but few have a clear read on results. We surveyed 100+ platform engineers on how they prioritize, measure, and manage cloud costs as AI reshapes their role. This session pairs those results with a practical look at how leading teams measure AI ROI, covering where you should spend your time, how much influence FinOps and the C-suite have over cost strategy, and how to apply unit economics to answer: was the AI investment worth it? You'll leave with best practices for avoiding AI bill shock and building ROI metrics for agentic workloads.

Speaker: Josh Glover
FinOps SME @ ProsperOps

Speaker: David Tepper
CEO and Co-founder @ Pay-i

Speaker: Grace Gui