Overview
Join us for our latest vLLM meetup, where we'll dive into the forefront of efficient large language model (LLM) inference and serving. As an open-source powerhouse for high-throughput, memory-efficient LLM deployments, vLLM continues to evolve rapidly, and this event is your opportunity to stay ahead of the curve.
Course Description & Learning Outcomes
Key Highlights
Introduction to vLLM & vLLM v1 Development Insights: Get an insightful introduction to vLLM from a major contributor, who will share key developments and experiences from vLLM v1.
Optimizing vLLM Inference on AMD Data Center GPUs: A Senior Product Manager for AMD AI Engineering will provide a deep dive into best practices and advanced techniques for maximizing vLLM inference performance on AMD's GPUs.
WEKA's Augmented Memory Grid (AMG): WEKA will present their innovative AMG, a high-performance, persistent key-value cache layer leveraging GPUDirect Storage (GDS) to deliver unparalleled performance for AI workloads.
Deploying AudioLLM with Ray: The MERaLiON team will share their experience and insights on deploying AudioLLM using Ray and VLLM, with a specific focus on effective autoscaling and load balancing strategies.
What You'll Walk Away With
By the end of this meetup, you'll have a deeper understanding of the evolving vLLM ecosystem, practical knowledge on integrating advanced solutions like AMG, and fresh ideas for enhancing your LLM serving infrastructure. This is more than a technical deep dive—it's a chance to network with fellow developers, maintainers, and industry leaders, fostering collaborations that drive open-source innovation forward.
Don't miss out—RSVP today and be part of the conversation shaping the future of efficient AI inference!
Schedule
Date: 27 Aug 2025, Wednesday
Time: 6:00 PM - 8:30 PM (GMT +8:00) Kuala Lumpur, Singapore
Location: 32 Carpenter Street, 059911
Agenda
| Day/Time | Agenda Activity/Description |
|---|---|
| 6:00 PM - 6:10 PM | Registration & light refreshment |
| 6:10 PM - 6:40 PM | Introduction to vLLM & vLLM v1 Development Insights Join us for an insightful introduction to vLLM. This session will also feature a major vLLM contributor sharing their experiences and key developments from vLLM v1. |
| 6:40 PM - 7:00 PM | Optimizing vLLM Inference on AMD Data Center GPUs: Best Practices & Deep Dive Join a Senior Product Manager for AMD AI Engineering to explore advanced techniques for maximizing vLLM inference performance on AMD GPUs. |
| 7:00 PM - 7:30 PM | WEKA Presents: Augmented Memory Grid (AMG) Discover WEKA's AMG, a high-performance, persistent KV cache layer powered by GPUDirect Storage, designed to optimize your data workflows. |
| 7:30 PM - 7:50 PM | Deploying AudioLLM with Ray Learn about MERaLiON team's experience deploying AudioLLM using Ray and VLLM, with a focus on autoscaling and load balancing. |
| 7:50 PM - 8:30 PM | Q&A, Dinner & Networking Cap off the evening with an interactive Q&A session, followed by dinner and an excellent opportunity to network with fellow attendees and speakers. |
Skills Covered
PROFICIENCY LEVEL GUIDE
Beginner: Introduce the subject matter without the need to have any prerequisites.
Proficient: Requires learners to have prior knowledge of the subject.
Expert: Involves advanced and more complex understanding of the subject.
- Cloud Computing (Proficiency level: Proficient)
- Machine Learning (Proficiency level: Proficient)
- Computer Science (Proficiency level: Proficient)
- Deep Learning (Proficiency level: Proficient)
- DevOps (Proficiency level: Proficient)
- Software Development (Proficiency level: Proficient)
Partners
