DataFramed

Almost every AI agent demo lands in roughly the same place: it works most of the time, looks remarkable, and then fails in a way no one anticipated. Self-driving cars hit this wall a decade ago, and agents are running into it now. For data and AI teams, the question is no longer whether agents can complete a task — it's whether they can complete it reliably enough to remove the human reviewer. Which categories of work tolerate a 90% success rate? Which absolutely don't? And where should the next layer of guardrails sit?
Ruslan Salakhutdinov is a UPMC Professor of Computer Science at Carnegie Mellon University and one of Geoffrey Hinton's former PhD students. He has previously served as Director of AI Research at Apple and VP of Research in Generative AI at Meta. His research focuses on deep learning, reasoning, and AI agents.
In the episode, Richie and Russ explore the most exciting use cases of AI agents today, long horizon tasks, the credit assignment problem, multi-agent systems, designing reliable human-in-the-loop workflows, agent safety and guardrails, embodied and physical AI, lessons from self-driving cars, the difference between academia and industry, and much more.
Links Mentioned in the Show:
• Claude Code (Anthropic)
• Yutori
• Waymo
• Apple Project Titan
• DeepSeek-V3 Technical Report
• Kimi K2 Technical Report
• Connect with Ruslan: LinkedIn
• AI-Native Course: Intro to AI for Work
• Related Episode: AI Agents at Work: What Actually Breaks (and How to Fix It) with Danielle Crop
New to DataCamp?
Learn on the go using the DataCamp mobile app

Empower your business with world-class data and AI skills with DataCamp for business

The data field has changed shape faster than almost any other. The role that used to be a statistician became a data scientist, became an ML engineer, and is now morphing into AI engineer. Consulting firms are hiring fewer entry-level analysts and more vibe-coders who can ship AI systems to production. For data and AI professionals, this raises immediate questions. Which parts of the work are most exposed to automation, and which are not? Where should you invest your time? And which backgrounds are now producing the strongest hires, whether you are building a team or trying to join one?
Ben Zweig is the CEO and Co-Founder of Revelio Labs, where he leads the development of a universal HR database built on over a billion public employment profiles and more than 5 billion job postings. He holds a PhD in Economics from the CUNY Graduate Center and teaches Data Science and The Future of Work at NYU Stern. Before founding Revelio Labs, he managed Workforce Analytics projects in the IBM Chief Analytics Office and worked as a data scientist at an emerging-markets hedge fund. He is the author of Job Architecture: Building a Workforce Intelligence Taxonomy.
In the episode, Richie and Ben explore why hiring is a broken two-sided market, why jobs are bundles of tasks not skills, building universal taxonomies from billions of job postings, which data careers resist AI, advice for hiring data talent, when traditional NLP beats LLMs, and much more.
Links Mentioned in the Show:
Ben's book — Job Architecture: Building a Workforce Intelligence Taxonomy
Revelio Labs
O*NET — the US government occupational taxonomy Ben critiques
Baruch Lev — The End of Accounting
Haskel & Westlake — Capitalism Without Capital
Justified Posteriors podcast (Andrey Fradkin & Seth Benzell)
Connect with Ben: LinkedIn
AI-Native Course: Intro to AI for Work
Related Episode: Our Data Trends & Predictions for 2026 with Jonathan Cornelissen & Martijn Theuwissen

New to DataCamp?
Learn on the go using the DataCamp mobile app
Empower your business with world-class data and AI skills with DataCamp for business

Time series data is everywhere — from inventory systems and energy grids to financial planning and product demand. As data volumes grow, the old ways of building individual forecasting models simply don't scale. How do you forecast hundreds of thousands of products without spending months on manual modeling? How do you know when to trust automation and when to step in? And what does it actually take to produce forecasts that business stakeholders will act on?
Rami Krispin is Senior Director of Data Science and Engineering at Apple Finance, where he leads teams working at the intersection of statistical modeling, machine learning, and production forecasting. He is the author of Hands-On Time Series Analysis with R, an open-source contributor, Docker Captain, and instructor. He holds an MA in Applied Economics and an MS in Actuarial Mathematics from the University of Michigan, where he began his journey learning time series on DataCamp — before going on to build his own course there.
In the episode, Richie and Rami explore time series foundation models and the case for scaling, traditional versus modern forecasting approaches, feature engineering in the business world, backtesting and model selection, risk management in automated forecasting, communicating forecast uncertainty to stakeholders, the evolving role of data scientists as architects, and much more.
Links Mentioned in the Show:
Forecasting: Principles and Practice (Rob Hyndman)
Nixtla
skforecast
Prophet
Connect with Rami
AI-Native Course: Intro to AI for Work
Related Episode: Developing Better Predictive Models with Graph Transformers

New to DataCamp?
Learn on the go using the DataCamp mobile app
Empower your business with world-class data and AI skills with DataCamp for business

Cloud data platforms now offer hundreds of services, plus a growing menu of SQL, NoSQL, and open source options. Unified environments promise a simpler path, but the hard trade-offs—consistency versus scale, single-writer versus sharded, RPO/RTO targets—still matter. In daily work, you may be deciding between SQL Server, Postgres, and a globally distributed JSON store, while also asking AI tools to draft queries and spot issues. Should you still learn SQL if an agent can write it? How do you validate the intent, performance, and security of generated queries? And can monitoring agents actually reduce on-call pain without taking away needed control?
Shireesh is the CVP of Databases at Microsoft. He leads product management, engineering, and cloud operations for Azure Databases as well as App Development for Microsoft Fabric. The products in his team’s portfolio include Azure SQL Database (on-prem, Hybrid and Cloud), Azure Cosmos DB, Azure PostgreSQL, and Azure MySQL.\\n\\n
Previously, as the Senior Vice President at SingleStore, Shireesh was responsible for end-to-end engineering and product vision of the company. Before moving to SingleStore, Shireesh was a founding member of Cosmos DB, where he architected, designed, and directly contributed to multiple key pieces of the services.\\n\\n
Shireesh has 20+ years of experience on large scale, big data, scale-out, relational and schema agnostic distributed systems across SQL, Azure Cosmos DB and PostgreSQL/Citus.
In the episode, Richie and Shireesh explore how AI agents are reshaping data stacks, why unified platforms like Fabric matter, how semantic models and ontologies reduce confusion in metrics, SQL and NoSQL choices on Azure, Postgres to Cosmos DB with guidance for builders, and much more.
Links Mentioned in the Show:
Microsoft Fabric
Azure Cosmos DB
What is Azure SQL Database?
Connect with Shireesh
AI-Native Course: Intro to AI for Work
Related Episode: Six Skills Data Professionals Need To Succeed with Abhijit Bhaduri, Brand Evangelist & Former General Manager of Global L&D at Microsoft
Explore AI-Native Learning on DataCamp

New to DataCamp?
Learn on the go using the DataCamp mobile app

Empower your business with world-class data and AI skills with DataCamp for business

Decision intelligence is showing up across data and AI teams as companies move beyond dashboards to decisions made with context. Graphs, entity resolution, and better data products are becoming core tools as messy, siloed data meets stricter risk and compliance needs. In day-to-day work, this means linking “James,” “Jim,” and “Jamie” across systems, enriching records with third‑party sources, and pushing models where the data already lives in your lakehouse. How do you trust your customer counts? Which links in a graph matter, and which are noise? Can graph-based context reduce LLM hallucinations enough for regulated decisions with humans still in-loop.
Jamie Hutton is the Co-founder and Chief Technology Officer of Quantexa, where he leads the company’s global research and development organization in advancing its market-leading Decision Intelligence Platform. With over two decades of experience pioneering data-driven technologies, Jamie has been at the forefront of innovations that connect and unify data at scale to solve complex real-world challenges. He is the creator of dynamic Entity Resolution, a pioneering capability that has redefined how the world’s leading organizations transform raw data into trusted, decision-ready intelligence. This innovation enables enterprises to prepare their data for AI, uncover new revenue streams, and expose hidden connections in even the most sophisticated criminal networks. By providing the foundation for accurate, explainable, and actionable insights, Jamie’s work has empowered governments, financial institutions, and global enterprises to make faster, smarter, and more confident decisions.
Prior to co-founding Quantexa, Jamie held senior technology and analytics leadership roles at SAS and Detica, where he delivered mission-critical solutions for organizations operating in some of the most complex and high-stakes environments in the world. Jamie holds a First-Class master’s degree in computer engineering and is recognized as a leading authority in contextual analytics, data integration, and applied AI for mission-critical decision-making.
In the episode, Richie and Jamie explore decision intelligence beyond BI, entity resolution across siloed data, building context graphs for fraud, AML, credit risk, and growth, how graph analytics separates meaningful links from noise, graph-RAG for LLMs to cut hallucinations, human-in-the-loop workflows, and ways to start today, and much more.
Links Mentioned in the Show:
Quantexa
Dun & Bradstreet Data Enrichment
Connect with Jamie
AI-Native Course: Intro to AI for Work
Related Episode: How Optimization Powers Decision Intelligence with Duke Perrucci & Ed Klotz, CEO and Senior Mathematical Optimization Specialist at Gurobi Optimization
Explore AI-Native Learning on DataCamp

New to DataCamp?
Learn on the go using the DataCamp mobile app

Empower your business with world-class data and AI skills with DataCamp for business