Dynamic Neural Network Compression for Scalable AI Deployment with Aditya Challapally

Deploying cutting-edge AI systems requires large models that can dynamically adapt to changing hardware, bandwidth, and task demands. In this talk, Aditya Challapally, Applied Science Lead at Microsoft and Connected AI Group Lead at MIT Media Lab, will introduce Dynamic Neural Network Compression, a framework for real-time compression and decompression of deep learning models.

Attendees will gain practical insights into new techniques for scalable AI deployment across cloud and edge environments using this new system and file format, Automatic Decompression Instructions (.adi).

Aditya Challapally Bio

Aditya Challapally is an Applied Science Lead at Microsoft, where he focuses on building agent-driven systems, scalable reinforcement learning workflows, and model compression techniques for deploying large models on edge devices. He also leads the Connected AI group at the MIT Media Lab, where his research centers on edge-first AI architectures, decentralized coordination systems, and distributed learning.