Jean-Georges Perrin

Mind the gap. Between the producer who shipped the data and the consumer who needed it yesterday. Between the spreadsheet on someone's laptop and the dashboard in the boardroom. Between the AI demo that wowed everyone and the AI system that actually runs in production. This 30-minute session is a practitioner's tour of how data contracts and data products close those gaps, frugally, without the rip-and-replace theatre and without pretending your data landscape is a greenfield (it isn't, and that's fine). Expect open standards from the Linux Foundation's Bitol project (ODCS, ODPS, and a few cousins), a brownfield-friendly approach you can start on Monday, and exactly zero slides telling you to "be data-driven." You already know. Let's talk about how. Bring tea. Or coffee. We won't judge.

Forget the models. Unstructured, unowned, uncontracted data is where AI goes to underdeliver. Most Gen AI initiatives deliver zero ROI, and Gartner expects 40% of agentic AI programs to be canceled by 2027. The analogy writes itself. Lufthansa Cargo doesn't ship freight without a manifest, a container, and a contract between shipper and carrier. Yet we routinely ship data to AI with none of that. Data products are the containers. Data contracts are the manifests. Open standards like ODCS and ODPS, part of the Linux Foundation's Bitol project, serve as a shared language that makes them readable across the supply chain. You've invested in a lakehouse. But even the best Delta Lake architecture breaks down when AI agents can't verify data quality, trace lineage, or resolve ownership when something goes wrong. Your infrastructure is solid. The contract layer above it is missing. And organizations that close that gap are 3.4x more likely to put AI into production. We'll show how Lufthansa Cargo's Data Cargo Platform puts that principle into practice, unifying 100+ real-time sources into a governed foundation for operational intelligence, powered by the Actian Data Intelligence Platform,…

Everything's bigger in Texas, including the expectations we put on data. Yet for many teams, "data products" remain a buzzword rather than something concrete they can build, ship, and trust.

When done right, governance is a growth engine. In this talk, Jean-Georges “jgp” Perrin will show how data contracts bring precision, trust, and accountability into your data and AI pipelines—without creating bottlenecks. Using the Open Data Contract Standard (ODCS) from the Linux Foundation’s Bitol project, you’ll see how organizations can cut downstream defects, accelerate AI model onboarding, lower compliance risk, and reduce firefighting—often in just days.

In the quest for modern data excellence, Data Mesh has emerged as a transformative paradigm—often imagined as the “Holy Grail” of federated data management. But the journey to implementing Data Mesh principles is fraught with challenges: how do you ensure interoperability, scalability, and governance in a distributed ecosystem?