With AI agents and GPU acceleration at the forefront, data science is entering a new era of efficiency and innovation. How are AI copilots transforming the way data scientists code and solve problems? Are they a reliable partner or a source of new complexities? On the other hand, the move to GPU-accelerated data science tools is revolutionizing model training and experimentation. What does this mean for the future of data science workflows? Explore these cutting-edge developments and their impact on the industry. Jean-Francois got a PhD in machine learning in the previous millennium. Given the AI winter at the time, he worked for a while on mathematical optimization software as dev manager for CPLEX in a startup. He came back to Machine Learning when IBM acquired the startup. Since then he discovered Kaggle and became one of the best Kagglers in the world. He joined NVIDIA 5 years ago and leads the NVIDIA Kaggle Grandmaster team there. Chris Deotte is a senior data scientist at NVIDIA. Chris has a Ph.D. in computational science and mathematics with a thesis on optimizing parallel processing. Chris is a Kaggle 4x grandmaster. In the episode, Richie, Jean-Francois, and Chris explore the transformative role of AI agents in data science, the impact of GPU acceleration on workflows, the evolution of competitive data science techniques, the importance of model evaluation and communication skills, and the future of data science roles in an AI-driven world, and much more. Links Mentioned in the Show: NVIDIANVIDIA RapidsFew shot learningConnect with Jean-Francois on Linkedin and Kaggle and check out Chris on KaggleCourse: Winning a Kaggle Competition in PythonRelated Episode: Becoming a Kaggle GrandmasterSign up to attend RADAR: Skills Edition New to DataCamp? Learn on the go using the DataCamp mobile appEmpower your business with world-class data and AI skills with DataCamp for business

Oftentimes, Kaggle competitions are looked at as an excellent way for data scientists to sharpen their machine learning skills and become technically excellent. This begs the question, what are the hallmarks of high-performing Kaggle competitors? What makes a Kaggle Grand Master? Today’s guest, Jean-Francois Puget PhD, distinguished engineer at NVIDIA, has achieved this impressive feat three times.  Throughout the episode, Richie and Jean-Francois discuss his background and how he became a Kaggle Grandmaster. He shares his scientific approach to machine learning and how he uses this to consistently achieve high results in Kaggle competitions. Jean-Francois also discusses how NVIDIA employs nine Kaggle Grandmasters and how they use Kaggle experiments to breed innovation in solving their machine learning challenges. He expands on the toolkit he employs in solving Kaggle competitions, and how he has achieved 50X improvements in efficiencies using tools like RAPIDS.  Richie and Jean-Francois also delve into the difference between competitive data science on Kaggle and machine learning work in a real-world setting. They deep dive into the challenges of real-world machine learning, and how to resolve the ambiguities of using machine learning in production that data scientists don’t encounter in Kaggle competitions.