One paper, 15 minutes from the presenters, and the room takes the rest.
Paper goes out a week ahead, so everyone walks in having read it and the discussion starts at the disagreement.
Roy Nissim (AI Inference Lead, Red Hat - vLLM, llm-d) presents "How Far Can Disaggregation Go? Attention-FFN Disaggregation for MoE Serving" (May 2026) and leads the discussion. Shraddha Sridhar (AI chip design, NVIDIA) moderates.
This event is a part of #SFTechWeek—a week of events hosted by VCs and startups to bring together the tech ecosystem. Learn more at www.tech-week.com.
