Please join this CSAIL Alliances TechTalk with Qualcomm

AI is rapidly evolving from models that generate content to intelligent systems capable of perception, reasoning, learning, and action. Realizing such capabilities at scale requires more than building ever-larger models in the cloud; fundamental advances across model architectures, learning and inference algorithms, system optimization, and computing platforms are essential.

This talk introduces Qualcomm AI Research and our work aimed at advancing the foundations of AI while bringing increasingly capable intelligence to edge devices. Research directions span efficient generative AI, model compression and accelerated inference, multimodal intelligence, reasoning, personalization, and agentic AI, connecting fundamental, applied, and platform research across the AI stack. A central theme is the journey from research breakthroughs to real-world intelligence, exploring how algorithmic innovation and full-stack co-design can make powerful AI models practical under real-world compute, memory, latency, and energy constraints, and ultimately enable a new generation of intelligent experiences that operate wherever computing happens.