Cover Image for verl Happy Hour @ ICML Vancouver
Cover Image for verl Happy Hour @ ICML Vancouver
Private Event

verl Happy Hour @ ICML Vancouver

Hosted by Ziyun Wei & 4 others
Register to See Address
Vancouver, British Columbia
Registration
Event Full
If you’d like, you can join the waitlist.
Please click on the button below to join the waitlist. You will be notified if additional spots become available.
About Event

​* Please join the waitlist. We will screen and admit manually due to capacity reasons.

Join top researchers and engineers in reinforcement learning, reasoning systems, and open-source agent infrastructure at the verl Happy Hour, a vibrant social and technical gathering colocated with ICML 2025 in Vancouver!

verl is an RL framework for LLMs, with 250+ contributors and 10k+ stars on Github. Hosted by developers of the verl project, Bytedance Seed and Volcano Engine, this event brings together experts advancing the frontiers of reinforcement learning, LLM-based agents, and scalable AI systems. Whether you’re building open-source agents, working on training infrastructure, or exploring automated reasoning, this is the place to connect.

We're thrilled to feature a stellar lineup of speakers:

🔹 Ziyun Wei – Research Scientist at Bytedance Seed and contributor to the verl project. Ziyun focuses on designing scalable distributed reinforcement learning systems.

🔹 Jing Xiong – Ph.D. student at the University of Hong Kong and contributor to SGLang. Jing's work explores efficient LLM inference and automated theorem proving.

🔹 Beidi Chen – Beidi leads research on efficient and scalable AI systems, spanning LLM inference, model architecture, reinforcement learning, and training-time acceleration.

🔹 Yi Wu – Yi Wu is an assistant professor at the Institute for Interdisciplinary Information Sciences (IIIS), Tsinghua University and leading the RL Lab at Ant Research. His research focuses on reinforcement learning and multi-agent learning.

About verl:
verl is an open-source reinforcement learning framework initiated by Bytedance Seed, designed for efficient and flexible training of large-scale models, especially for LLMs. verl enables rapid prototyping while scaling seamlessly across thousands of GPUs. It has been adopted and contributed by leading research labs and companies to accelerate RL pipelines, including Amazon, LinkedIn, Alibaba Qwen, Xiaomi, RedNote, etc

About Volcano Engine:
Volcano Engine is ByteDance's cloud and AI infrastructure platform, a top provider in the AI cloud market. Backed by the Volcano Ark architecture, it provide scalable and high performance services for ByteDance's advanced models, including the Doubao 1.6 LLM series, video model Seedance 1.0 Pro for video generation, etc.

Welcome to the party - see you there!

* the location is near the Vancouver Convention Centre, which will be shared to approved participants due to capacity reason.

* this event is on-site only, no recording available. If you cannot make it, please cancel registration.

Location
Please register to see the exact location of this event.
Vancouver, British Columbia