The Data Culture Podcast | Podcast in RSS.com

The Data Culture Podcast

The Data Culture Podcast

di Sid Atkinson and Lee Harper

Amministrazione Tecnologia

Stagione 1

Ai Exposes Our Untended Gardens

Culture eats strategy for lunch, and on this episode it eats your AI roadmap too. Sid Atkinson sits down with Mathias Vercauteren to make the case that data is the fourth production factor, AI is the fifth, and most companies are managing both with a fraction of the rigor they apply to money and people. They get into why Amazon runs circles around competitors who sit on far more data, how shadow AI quietly outgrew the old BYOD headache, and what the EU AI Act signals for everyone operating outside Europe. Mathias also lays out where to actually begin on Tuesday morning, and the answer is refreshingly unglamorous: start counting what you've got.

Forward Deployment Engineers, Integration Humps, and the Two Last Miles

The term "forward deployment engineer" is everywhere right now, Palantir popularized it, Anthropic is leaning into it, and every data and AI platform seems to be packaging it as their answer to the last mile problem. But what is it, really? Michael Wharton, VP of Engineering at Kung Fu AI, joins Sid and Lee to cut through the noise. They dig into where the FDE model works, where it falls short, why process matters more than the individual, and why buying the platform and buying the deployment help are two very distinct purchases. They also get into token maxxing versus quality, the bimodal split happening across engineering teams right now, and why the speedup from agentic coding tools is probably more modest than most leaders think.

Socks, Crocs, and AI Security

As a founder of the Berryville Institute of Machine Learning, Gary McGraw has been researching AI security since before most people knew what machine learning was. He's identified 78 risks across ML systems and was sounding the alarm on recursive pollution and model collapse long before those terms went mainstream. He joins Sid and Lee to break down what practitioners need to understand about the systems they're implementing, why 23 of those risks live in a black box controlled entirely by the foundation model vendors, and what good governance looks like when you can't see inside the thing you're governing.

Librarians, Lawyers, and Judges: The Future of Data Work

Julia Bardmesser, CEO of Data4Real and former Chief Data Officer at Voya Financial, brings 25+ years of financial services experience to a conversation about what's actually standing between companies and real AI value. She and Sid dig into why AI layered on top of broken data and workflows delivers little ROI, why the unsexy work of semantic clarity and data management at scale is now more critical than ever, and why the human roles of the future might look less like engineers and more like librarians, lawyers, and judges.

Agent to Agent to Human to Agent: Are you there Claude? It's me Margaret

You've heard of UX. You've heard of developer experience. Now meet AX — Agent Experience — and it's about to reshape how enterprises think about their entire tech stack. Lee Harper talks with Vihan Patel of Australia's Mantel Group about what happens when AI agents start bumping up against APIs that were never designed for them, why e-commerce is ground zero for the agentic shift, and the uncomfortable truth that enabling Cursor for your whole dev team doesn't automatically make you faster. They also tackle the governance nightmare of scaling from two agents to twenty, and whether we're heading toward a world where we get paid for decisions, not deliverables.

Behavior Change Over Better Tools: The Real Work of Data Governance

Most data programs don't fail because of bad technology, they fail because nobody wants to talk about the hard stuff. John Ladley has spent decades proving that behavior change is what separates the programs that stick from the ones that quietly die. In this episode, he's not holding back. From walking away from clients who weren't serious to telling a CIO to admit he didn't know what he was doing, John shares the unfiltered lessons that most consultants won't say out loud. If you work in data and you're tired of watching good ideas go nowhere, this one's for you.

Becoming a Data Catalyst: The Spark Your Data Culture Needs

Data governance is often seen as restrictive—but what if it’s actually the key to making data work? Bob Seiner joins Sid Atkinson to challenge how organizations think about governance, accountability, and behavior change. Drawing from decades of experience, Bob explains why governance efforts stall, what shifts when organizations focus on outcomes instead of activity, and how ownership, not tooling, drives real adoption. He also introduces the concept of data fluency, exploring why the ability to communicate with data is becoming just as important as understanding it.

Data Culture Isn't a Mystery

Hosts Sid Atkinson and Lee Harper engage with guests Gary Griffin and David Holcomb, authors of "Building a Data Culture: The Usage and Flow Data Culture Model", to explore the concept of data culture. They challenge the notion that culture is unmanageable, presenting frameworks and models that make culture visible and actionable. The discussion covers the three R's of data culture, the usage and flow model, the significance of microcultures, and the differences between public and private sector dynamics. The conversation culminates in the introduction of the Data Culture Institute, aimed at equipping leaders to build and sustain a data-driven culture.

Semantic Patches: the Inmon/Kimball Debate of the LLM Era

Data management purity and pragmatism built opposed camps starting in the 1990s, and echoes of those tenants present themselves today in Generative AI. Ontology purists and vector-based engineers appear at opposite ends, but how might we reframe so everyone's in the same picture? Tracy Talbot is a consummate data practitioner and has seen it all—from mainframes to machine learning. In this episode, she sits down with Sid Atkinson to reveal how the age-old Kimball vs. Inman debates mirror today’s AI struggles—and why her concept of “semantic patches” could be the key to keeping generative AI grounded in reality.

From Team of One to Data Trust: Turning Government 'No's into 'Yes's

Building a data program isn't all about the tech—it's about getting people to actually work together. Carlos Rivero shares what it was really like starting as Virginia's first Chief Data Officer with no team, no budget, and a lot of skeptical agencies. He talks about how he learned to be more of a "people wrangler" than a data expert, turning critics into allies and figuring out that most of the job was just getting everyone on the same page.

1 di 7