Date: 21st October 2025
Location: San Francisco, California
Kickoff at the PyTorch Conference
Whether you’re attending the conference or simply based in the SF Bay Area, join us as we dive into the future of AI infrastructure and its role in advancing reinforcement learning, large-scale inference, and media/world models.
At this session, the DataCrunch team, alongside leading frontier AI labs (TBA), will share hard-won insights from building and scaling systems at the edge of what’s possible. You’ll also get an exclusive first look at the B300 and GB300 NVL-72 systems, and a glimpse of what’s next in AI infrastructure.
Expect deep technical discussions, practical learnings from practitioners, and the chance to connect with like-minded engineers. Stick around to enjoy food, drinks, and sharp conversations.
Speakers
- Paul Chang - ML Engineer at DataCrunch
- Erik Schultheis - Postdoctoral Researcher at IST Autria
Agenda
-
5:30pm – Arrival
-
6:00pm – Talks + Q&A
-
7:00pm – Networking, food, & drinks
-
9:00pm – Wrap-up
Note: Pacific Time (GMT-7)
Talks
Talk 1: Training world models using B200s by Paul Chang
Talk 2: Training quantized LLMs efficiently on consumer GPUs by Erik Schultheis
Who Should Join?
-
AI researchers
-
ML engineers
-
Technical founders
-
AI product managers
This event is for those staying ahead of the curve with AI infra, optimization techniques, and production-grade systems at scale.
Other hosts and speakers (TBA)
→ Sign up here
