vLLM Meetup HK
.jpg?bc=ffffff&h=290&w=405&rev=a94836f5cf5f4fae887ab4a39bc312e0&hash=2364BD155194D6FE9A1D5E15ECE44DCA)
vLLM Meetup HK
COMP was proud to co-organise the inaugural vLLM Meetup HK with Red Hat Inc. on 7 March 2026. The event served as a vibrant hub for the local AI community, drawing an overwhelming response with over 400 registrations for the 150-seat capacity venue.
The meetup brought together a diverse group of engineers, researchers, and industry leaders from some of the world’s most influential organisations in AI and large language models (LLMs), including NVIDIA, AMD, MetaX, Bytedance, Tencent, Minimax, and Microsoft. These experts converged at PolyU to exchange ideas and explore the future of LLM inference, with a particular focus on emerging trends in omni-modality and performance optimisation.
Attendees benefited from a comprehensive program that bridged theory and practice. The event provided participants with practical insights into LLM inference through a series of talks and hands-on sessions. Furthermore, guests had the unique opportunity to learn about state-of-the-art LLM inference development directly from industry leaders and core contributors to the vLLM project, while also networking with the wider open-source community to foster future collaboration.
The meetup brought together a diverse group of engineers, researchers, and industry leaders from some of the world’s most influential organisations in AI and large language models (LLMs), including NVIDIA, AMD, MetaX, Bytedance, Tencent, Minimax, and Microsoft. These experts converged at PolyU to exchange ideas and explore the future of LLM inference, with a particular focus on emerging trends in omni-modality and performance optimisation.
Attendees benefited from a comprehensive program that bridged theory and practice. The event provided participants with practical insights into LLM inference through a series of talks and hands-on sessions. Furthermore, guests had the unique opportunity to learn about state-of-the-art LLM inference development directly from industry leaders and core contributors to the vLLM project, while also networking with the wider open-source community to foster future collaboration.
