Joo-Young Kim is the CEO of HyperAccel and an Associate Professor at KAIST, where his research focuses on specialized AI hardware and large language model inference. ~~~~~~~~~~~~~This episode is brought to you by Nebius — the ultimate cloud for AI innovators.Nebius provides AI infrastructure you can count on, combining reliability and speed with flexibility and engineering support unmatched by hyperscalers.AI leaders like Meta, Shopify, and Higgsfield already partner with Nebius to run their AI workloads. Plus, venture-backed startups can save up to $150,000 on compute costs when they apply for access. Visit nebius.com or nebius.com/startups to learn more~~~~~~~~~~~~~He founded HyperAccel in 2023 after spending more than seven years at Microsoft, contributing to the Catapult project, which developed FPGA accelerators for cloud data centers. His work included accelerating workloads such as data compression, encryption, video transcoding, and machine learning across Microsoft’s cloud infrastructure. At HyperAccel, he is developing the LLM Processing Unit (LPU), a specialized AI processor designed to run large language models with lower latency and improved cost and power efficiency. His career bridges academic research, hyperscale cloud hardware, and semiconductor entrepreneurship, with a focus on addressing the infrastructure bottlenecks that emerge as generative AI scales.Topics
Why large language model inference requires a different approach to AI chip architecture
Designing specialized AI semiconductors to improve inference speed, energy efficiency, and operating costs
Turning academic research and hyperscale cloud experience into a semiconductor startup
Podden och tillhörande omslagsbild på den här sidan tillhör
Grace Gong. Innehållet i podden är skapat av Grace Gong och inte av,
eller tillsammans med, Poddtoppen.
