Bridging AI Research and Real-World Impact — Dr. Petar Veličković, Senior Staff Research Scientist at Google DeepMind

Learn more about how CambridgeSpark.com is helping organisations and professionals master Data & AI skills to stay ahead in the digital era. In this episode of Data & AI Mastery, host Dr. Raoul-Gabriel Urma sits down with Dr. Petar Veličković, Senior Staff Research Scientist at Google DeepMind and Affiliate Lecturer at the University of Cambridge. Together they explore how cutting-edge AI research transitions from theory to practice, spanning from breakthroughs in graph neural networks that power Google Maps worldwide, to the use of AI as a discovery partner in mathematics. Petar shares his journey from Serbia to Cambridge to DeepMind, and the pivotal lessons learned at every stage. Listeners will also hear how DeepMind’s graph machine learning models improve global travel-time predictions in Google Maps, how AI can accelerate scientific discovery, including mathematical proofs and what the concept of an “AI scientist” might mean for the future of research. Whether you’re a data scientist, ML engineer, or executive exploring AI strategy, this episode offers a masterclass in applying AI innovation with purpose and impact. Be sure to follow Data & AI Mastery wherever you listen to your podcasts to never miss an episode. Chapter Markers: (06:00) - Inside DeepMind; what a Senior Staff Research Scientist does (13:00) - From research to real-world impact: challenges & latency trade-offs (19:00) - AI for mathematics; collaborating with top researchers (25:00) - Defining the “AI scientist” and the future of autonomous research (30:00) - What’s next for AI agents and self-directed systems (31:00) - Quick-fire round; programming languages, school favourites & music (33:00) - Raoul’s takeaways; user focus, AI as a thought partner, safe deployment Useful Links: Connect with Petar on LinkedIn Follow Raoul for more AI insights on LinkedIn Explore Cambridge Spark’s AI upskilling programmes at cambridgespark.com