Dynamo Day

Dynamo Day brings together the open-source Dynamo community for a day of learning, sharing, and connection. Whether you’re a longtime contributor, a builder using Dynamo in production, or just getting started, this event is designed to help you learn from each other and shape what’s next.

Why Attend Dynamo Day?

Dynamo Day is about the people behind the project—the users, builders, maintainers, and partners who are building real solutions with Dynamo every day.

At Dynamo Day, you'll:

  • Learn about the latest updates and innovations in Dynamo
  • Hear real-world success stories and lessons learned from the community
  • Discover best practices for building, deploying, and scaling with Dynamo
  • See how users and partners are building products and integrations on top of Dynamo
  • Connect with others who care about the future of the Dynamo Ecosystem

What to Expect

Community Stories

Talks and demos from community members using Dynamo in production

Partner Perspectives

Insights from partners building solutions and extensions on Dynamo

Maintainer Insights

Opportunities to learn directly from maintainers and contributors

Best Practices

Practical guidance for building, deploying, and scaling with Dynamo

Who Should Attend?

Dynamo Day is open to anyone interested in Dynamo, including: 

  • Developers and engineers building with Dynamo
  • Open-source contributors and maintainers
  • Platform and infrastructure teams
  • Architects and technical leaders
  • Partners and ecosystem builders
  • Anyone exploring Dynamo for the first time

All experience levels are welcome.

Dynamo Ecosystem

Agenda

8:00 a.m.–8:15 a.m.
Opening Remarks and Inference Introduction

Why Thinking Models Demand Codesign

Amr Elmeleegy, Product Marketing, NVIDIA

8:15 a.m.–8:30 a.m.
State of Dynamo

Product Progress, Community Growth, and What’s Ahead

Harry Kim, Product Manager, NVIDIA

8:30 a.m.–9:00 a.m.
Under the Hood: Dynamo Inference

From Architecture to Performance

Neelay Shah, Principle Engineer, NVIDIA

Kyle Krannen, Engineering Lead, NVIDIA

9:00 a.m.–9:30 a.m.
Learnings From Dynamo Users

Field Experiences and Lessons Learned

Philip Kiely, Head of Developer Relations, Baseten

Ameen Patel, Member of Technical Staff, Prime Intellect

Lei Pan, Software Engineer, Pinterest

Salina Wu, Software Engineer, Pinterest

10:00 a.m.–10:15 a.m.
Inference: Today and Tomorrow

Fireside Chat With Google

Ling Lin, Software Engineer, Google

Sean Horgan, Product Manager, Google

Anish Maddipoti, Product Manager, NVIDIA

10:15 a.m.–11:00 a.m.
Inference OSS Ecosystem

Framework Updates and Integration Insights

Simon Mo, Lead Developer, vLLM

Baizhou Zhang, Core Developer, SGLang

Laikh Tewari, TensorRT-LLM Product Manager, NVIDIA

Carlos Costa, Member of Technical Staff, IBM

11:30 a.m.–11:45 a.m.
Building Inference Together

Perspectives From Microsoft’s Contribution to Dynamo

Sertaç Özercan, Principal Software Engineering Manager, Microsoft
11:45 a.m.–12:00 p.m.
Solving KV Caching Bottlenecks

Techniques to Optimize Memory and Latency

Yihua Cheng, CTO and Cofounder, Tensormesh

Akshatha Kamath, Product Manager, NVIDIA

12:30 p.m.–1:00 p.m.
The Future of AI Inference

Expert Perspectives From University of California San Diego

Hao Zhang, Assistant Professor, UC San Diego

Time Zone: (UTC-08:00) Pacific Time (US & Canada) [Change Time Zone]

Highlighted Resources

How NVIDIA GB200 NVL72 and NVIDIA Dynamo Boost Inference Performance for MoE Models

NVIDIA Dynamo optimizes MoE LLM inference through disaggregation, boosting efficiency, throughput, and GPU utilization.

How to Reduce KV Cache Bottlenecks With NVIDIA Dynamo

NVIDIA Dynamo enables KV cache offloading to reduce inference costs and improve LLM user experience.

Speakers

Hao Zhang

Hao Zhang

UC San Diego

Assistant Professor

Phillip Kiely

Phillip Kiely

Baseten

Head of DevRel

Sertaç Özercan

Sertaç Özercan

Microsoft

Principal Software Engineering Manager

Simon Mo

Simon Mo

vLLM

Lead Developer

Yihua Cheng

Yihua Cheng

Tensormesh

CTO and Cofounder

Baizhou Zhang

Baizhou Zhang

SGLang

Core Developer

Ameen Patell

Ameen Patell

Prime Intellect

Member of Technical Staff

Ling Lin

Ling Lin

Google

Software Engineer

Carlos Costa

Carlos Costa

llm-d

Member of Technical Staff

Lei Pan

Lei Pan

Pinterest

Software Engineer

Virtual Event Details

This event will be livestreamed so anyone globally can tune in

Thursday, January 22, 2026, 8:00 AM - 1:00 PM