Modern semiconductor fabs face increasing complexity due to shrinking feature sizes, advanced packaging (e.g., chiplets), diverse product portfolios, and massive data volumes. To remain competitive, fabs require sophisticated tools for accurate prediction and optimized resource allocation, as traditional static models and heuristics struggle with the dynamic and unpredictable nature of high-volume manufacturing. This joint presentation from minds.ai and Micron introduces a multi-agent Reinforcement Learning (RL) workflow designed for holistic fab optimization, integrating advanced predictive capabilities with intelligent, adaptive agentic control.
.jpg "Sumit Sanyal photo")