Dive into a 50-minute podcast episode featuring Erik Bernhardsson, CEO & Founder of Modal Labs, as he discusses the future of machine learning infrastructure with host Lukas Biewald. Explore how Modal is enhancing developer experience, handling large-scale GPU workloads, and simplifying cloud execution for data teams. Gain valuable insights on AI, data pipelines, and building robust ML systems. Learn about the evolving roles in ML engineering, the growing need for GPU access, challenges in scaling ML workloads, and the importance of developer experience in building ML platforms. Discover how Modal optimizes AI inference, supports multiple programming languages, and leverages open source in machine learning infrastructure. Reflect on the future of ML, including the debate between training custom models and using pre-built ones.
Elevating ML Infrastructure - Future of Machine Learning Platforms
Overview
Syllabus
– Introduction: Lukas introduces Erik Bernhardsson, CEO & Founder of Modal Labs
– What Modal Labs Does and Its Vision for ML Infrastructure
– Importance of Developer Experience in Building ML Platforms
– Evolving Roles: From Data Teams to Machine Learning Engineers
– The Growing Need for GPU Access and Cloud Infrastructure
– Challenges of Scaling ML Workloads in Production
– Prioritizing Features and the Development Process at Modal
– How Modal Optimizes AI Inference and Custom Workflows
– Thinking Beyond Python: Supporting Multiple Languages in ML
– The Role of Open Source in Machine Learning Infrastructure
– Future of ML: Training Custom Models vs. Using Prebuilt Ones
– Conclusion
Taught by
Weights & Biases
