Overview
This unique event is crafted for developers, data scientists, AI engineers, and enthusiasts eager to explore and demystify advanced Large Language Model (LLM) inference techniques. Whether you’re an expert or just starting your AI journey, you’ll find valuable insights and practical guidance tailored to your experience level.
Detailed Agenda
🌅 Morning Agenda
9:00 AM – 9:30 AM
- Doors Open and Check-In Registration
- Morning networking opportunity: connect with fellow developers over coffee.
9:30 AM – 10:00 AM
- Welcome Address
- Introduction from SGInnovate and the event organizing committee (Embedded LLM).
- Insights into the vision and future direction of the vLLM Asia Community.
10:00 AM – 10:30 AM
- Networking Break
- Continue conversations with speakers and peers over refreshments.
10:30 AM – 12:00 PM
- vLLM Updates and Technical Talks
- 10:30 – 11:00 AM: State of vLLM – Current Status & Future Roadmap by Chen Zhang and Cyrus Leung.
- 11:00 – 11:30 AM: Running vLLM in Production by Tun Jian Tan.
- 11:30 – 11:40 AM: Open Q&A with vLLM Experts (Chen Zhang, Cyrus Leung, Tun Jian Tan).
- 11:40 AM – 12:00 PM: Roundtable Panel: “Infrastructure Evolution for AI Inference” featuring experts from A*STAR-MERaLiON, AI Singapore SEA-LION, and SeaAI-Sailor2.
12:00 PM – 1:00 PM
- Lunch and Social Hour
- Informal interaction, networking, and collaboration opportunities.
☀️ Afternoon Agenda
1:00 PM – 1:30 PM
- Doors Open and Check-In Registration (Afternoon Session)
1:30 PM – 3:30 PM
- Technical Talks and Deep-Dive
- 1:30 – 2:30 PM: AMD AI Software Introduction and LLMs Optimization with vLLM for AMD GPUs by George Wang.
- 2:30 – 3:00 PM: DeepSeek R1 Inference Optimization Case Study by Bruce Xue.
- 3:00 – 3:30 PM: vLLM ROCm New Features by Haichen Zhang.
3:30 PM – 4:00 PM
- Networking Break
- Engage and share insights with speakers and fellow participants.
4:00 PM – 6:30 PM
- Hands-On Technical Workshop: From Zero to Production
- Deploy Optimized LLMs with vLLM on AMD (FREE AMD GPU access provided during the workshop)
- Build GenAI use cases on JamAI Base
- Walk away with practical deployment examples applicable to your projects.
- Conducted by Tun Jian Tan, Dr. Pin Siang Tan, and Bruce Xue.
🌆 Evening Agenda
6:30 PM – 9:30 PM
- Social Mixer and Networking Evening
- Opportunity to mingle, share experiences, and enjoy dinner.
6:30 – 7:30 PM: Dinner & Social Hour
7:30 – 7:40 PM: Introduction to Evening Lightning Talks
-
7:40 – 8:00 PM:
MERaLION – AudioLLM
Advancing Multimodal Speech-Text Foundation Models for Singapore and SE Asia
Presented by Yingxu He (Senior Research Engineer, A*STAR). -
8:00 – 8:20 PM:
Making LLMs Reason Better: GRPO + vLLM Journey
Presented by Dr. Ye Hur Cheong (Principal LLM Solutions Architect, Embedded LLM). -
8:20 – 8:40 PM:
Speculative Decoding for Multilingual LLM
Presented by Cunxiao Du (Researcher, SEA AI Lab).
8:40 – 9:30 PM: Wrapping up, Networking, and Event Close.
🚀 Speakers Highlights
Our distinguished lineup includes:
- Chen Zhang: PhD Student at Tsinghua University, vLLM Committer.
- Cyrus Leung: PhD Candidate at HKUST, vLLM Maintainer.
- Tun Jian Tan: Principal Engineer, key vLLM Committer, AMD inference optimization expert.
- George Wang: Director, AI Software Product Engineering, AMD.
- Bruce Xue: AI Product Application Engineer, AMD.
- Haichen Zhang: Senior PM, AMD AI Product Marketing.
- Dr. Pin Siang Tan: Co-Founder & CTO, Embedded LLM.
- Yingxu He: Senior Research Engineer, I2R, A*STAR.
- Dr. Ye Hur Cheong: Principal LLM Solutions Architect, Embedded LLM.
- Cunxiao Du: Researcher, SEA AI Lab.
📍 Venue & Time
- Date: April 3, 2025 (Thursday)
- Time: 9:00 AM – 9:30 PM (Full-day event)
- Venue: SGInnovate Office