2026 Sessions

AI's productivity surge masks a complex reality: it shifts developer friction from writing boilerplate to a new "verification tax" of auditing machine-generated code. DORA research reveals AI is an amplifier. It magnifies the strengths of high-performing teams, but in chaotic environments, it accelerates technical debt and cognitive load. We will explore the AI adoption "J-Curve" to understand why early dips in flow and stability are expected. Using the DORA AI Capabilities Model, we'll show how quality internal platforms and healthy data ecosystems empower developers to focus on meaningful work. Move beyond the hype of raw code generation. Learn to clear systemic roadblocks and build an environment where AI truly amplifies developer experience, creativity, and flow.

Blackstone's AI journey started with a pretty ordinary goal: get new developers up to speed faster. Then coding agents showed up, and the problem got a lot more interesting. Because in a regulated industry, "move fast" and "get compliance to trust this" don't usually go together. This session walks through how that tension actually got resolved, including a decision that raised some eyebrows internally: coding agents run only in cloud-based dev environments, never on endpoints. It's a tradeoff, and Blackstone will talk through why they made it. The most interesting part might be the data. When Blackstone lined up GitLab activity against AI tool usage, the productivity story wasn't what anyone expected going in. You'll walk away with a real framework for getting compliance on board with AI agents, an honest look at the endpoint tradeoff, and a better way to measure whether AI is actually helping, not just being used.

As agents absorb more of the execution across the SDLC, every service that ships faster becomes another unit to own and secure, and aggregate risk compounds well ahead of any single team's ability to see it. Operational Excellence is how engineering leaders keep the system accountable: measuring where the org actually stands, catching risk before it becomes an incident, and moving people and attention to where they matter most. This panel brings together leaders who run recurring operational reviews and treat their organization as something they can observe and improve, not just react to. Sample questions: What does an operational excellence review look like in practice, how often should you run one, and who should be in the room? Which signals tell you output is outpacing your ability to operate it safely? How do you decide where to spend engineering time when everything looks urgent? What breaks first when velocity jumps 10x, and how do you catch it early?

Migrations and modernization programs are often slow, political, and difficult to finish. AI-accelerated throughput raises the stakes: more services, more dependencies, and more code shipping before anyone agrees on what "ready for production" means. This panel examines how leaders turn production readiness into a standard instead of a launch-day scramble, keep large migrations on track across dozens of teams, and modernize legacy systems without freezing delivery. Panelists share what worked, what stalled, and what they would do differently. Sample questions: How do you define "production-ready" so it holds as a standard, not a one-time checklist? What makes a migration actually finish instead of stalling at 80%? How do you keep legacy modernization moving without freezing feature work? Where does AI-generated code create readiness gaps you didn't have before?

When our platform crossed 5 million artifacts, the real challenge wasn't just storage volume, it was the operational complexity that came with it: too many repositories, frequent corruptions, inconsistent patterns, rising storage costs, and security scanning that produced more noise than action. In this session, I'll share our practical journey to building a leaner artifact delivery model using JFrog Artifactory with a focus on minimal configuration, controlled spend, and better developer experience. We'll cover how we simplified our repository strategy, standardized lifecycle flows, reduced custom configuration, and introduced retention discipline, by aligning them to release decisions instead of just reporting dashboards. This is not a "perfect architecture" talk, it's a field-tested story of what worked, what didn't, and the trade-offs we had to make to scale responsibly.

AI-augmented development has not-so quietly become one of the fastest-growing line items in the engineering budget: model inference, agent fleets, copilot seats, GPU and accelerator spend. Most organizations still can't answer a basic CFO question: what did we get for it? Spend is approved on faith and scrutinized a quarter later, because the instrumentation to measure return hasn't kept pace with how fast teams ship AI. "Do more with less" and "ship AI everywhere" pull in opposite directions, and reconciling them lands on engineering and platform leaders. The real cost is bigger than the token bill: AI-generated code carries a downstream load of review, security, and maintenance that few teams price in. This roundtable convenes leaders to compare how they turn that spend into measurable efficiency: attributing cost to the teams and features that drive it, cutting the waste hiding in idle capacity and redundant calls, and knowing whether or not the investment was worth it.

Careers are made, and ruined, on migrations. The technology is often the easy part. The hard part is the people, the stakeholders, the politics, and the business pressure that doesn't pause while you're mid-execution. This roundtable is a peer conversation for engineering and platform leaders who've been in the fire. We'll go beyond the technical playbook to explore how the best leaders navigate the human and organizational complexity of migrations that truly matter. Where success or failure defines the business, and sometimes your career along with it. Come ready to share what you've learned, what you'd do differently, and what advice you wish you'd had.

Every engineering org wrestles with the same tension: enough standardization to move fast, but not so much that teams feel constrained. In this roundtable, we'll explore how platform teams can use service templates, scaffolds, and golden paths to encode best practices without becoming a bottleneck. Joshua Marble, Senior Manager of DevX at Archer, will share how his team uses Cortex alongside opinionated templates to reduce the cognitive load on development teams and open the floor to discuss what's worked (and what hasn't) across our organizations.

We have had to go from zero internal support to a 24/7 on-call support operation through a combination of a new SRE team, a best-in-class incident management solution (incident.io) and a central service catalog (Cortex). This talk will outline how we partnered with our vendors to implement a slick incident management operation, the pain points we solved and our future direction augmented by AI.

There are two approaches being taken by most orgs, either together or in a divergent manner. This round table discussion and deep dive will cover two areas to unpack the positives and negatives of both approaches. Organisations which are pursuing software factories - dark or guided - where the emphasis is on building the machine which builds the machine enabling non-engineers to ship and deliver new capabilities. Software Engineers as "pilots" where humans stay in the driving seat and delivery is augmented and magnified by Agents. Are both approaches valid and how do they interact? What has worked, and what has failed at your organisation?

Vibe coding doesn't stop at the edge of your platform team - agents are writing, committing, and shipping code faster than anyone can review it, and half of them (if you're lucky) are doing it with zero adult supervision. Standing firm means a shield wall; cost caps that don't bankrupt you, approval gates that don't grind delivery to a halt, and scanning pipelines that catch the mess before it reaches prod. This roundtable trades war stories and real-world solutions on building that governance layer — for developers and citizen developers alike — without turning your platform team into the AI police.

Find out whether your AI output is creating value or just accumulating risk. Led by Taylor Schmidt, Director of Customer Education and Delivery, Cortex. Your teams are shipping more code than ever, and most of the frameworks meant to measure the success of an engineering organization look at developer productivity. They fail to measure whether accelerated output is landing as reliable, secure software or piling up as invisible risk and tech debt. DRIVE measures organizational effectiveness in the age of AI, across five pillars covering Delivery, Reliability, Initiatives, Vigilance, and Efficiency. This certification workshop takes you through each pillar, how the signals fit together, and how DRIVE complements the DORA metrics you likely already track rather than replacing them. By the end you'll be able to: Assess where your organization stands on each of the five DRIVE pillars Explain each pillar and its signals to the engineers and executives you report to and with Walk out DRIVE-certified, with a plan for the metrics you need to track your org's maturity going forward No prior DRIVE or Cortex experience needed. Come with a rough picture of how your org ships today and you'll leave with a structured way to measure it.

At the leadership level, the problem is rarely too little data. It's parsing signal from noise. It's parsing signal from noise: deploy frequency, MTTR vulnerability counts scattered across teams. By the time someone's stitched it together by hand, you're staring at a wall of numbers, unsure what's changed or where to focus. As agents push more code, the volume grows. This workshop walks you through the Operational Excellence review, using DRIVE, a framework built around five pillars: Delivery, Reliability, Initiatives, Vigilance, and Efficiency. You'll work with real data in Cortex and learn how to run the review itself, end to end. You'll leave with: A repeatable OpEx review agenda for your leadership team Which DRIVE signals belong in that room, and which to cut A way to turn the review into clear & owned action items Bring a laptop. You'll work directly in Cortex.