Practical AI Federated learning in production (part 2)
26 snips
Jun 4, 2025 
Chong Shen, a research engineer at Flower Labs with a background in computational physics, shares insights on building production-ready federated learning systems. He discusses the user-friendly Flower framework, essential for handling sensitive data across industries. Topics include the design challenges for developers, the integration of diverse models, and how the generative AI boom is influencing future developments. The conversation sheds light on the balance between usability and production demands, especially in sectors like healthcare.
AI Snips
Chapters
Transcript
Episode notes
Chong Shen's Career Path
- Chong Shen pivoted from computational physics academia to industry through consulting in federated learning.
- He joined Flower Labs driven by his passion for open source and distributed learning frameworks.
Federated Learning Basics
- Federated learning trains models locally where data is generated instead of moving data to a central location.
- Model weights are aggregated centrally to build a global model without data sharing.
What Drives Federated Learning Adoption
- The primary driver into federated learning is the inability or unwillingness to share data across parties.
- Another driver is coordinating multiple datasets within one company too complex for centralized training.