On April 3, 2025, the Embedded LLM team, in partnership with SGInnovate, proudly hosted the first-ever vLLM Asia Developer Day at SGInnovate’s office in Singapore. The event brought together a vibrant community of developers, data scientists, AI engineers, and industry leaders to explore the latest in LLM inference technologies.
We’re thrilled to share that the event was a resounding success, with over 100 participants joining us for a full day of technical deep dives, workshops, and networking.
🔑 Event Highlights
Morning Sessions: Panels & Workshops
The day kicked off with a successful LLM tech lineup with a vision for the growing vLLM Asia community.
Technical Updates
- Chen Zhang and Xiyou Liang presented the State of vLLM
- AMD team shared practical insights on running vLLM in production
- Smart Panel
- Leaders from APTAR BPEA-JM, AI Singapore SEA Labs, and Noah Robert explored how AI inference infrastructure is evolving in APAC
Afternoon Sessions: Opportunities & Engineering
Deep Technical Talks
- With George Sung, Jiayi Shen, and Haotian Zhang highlighted LLM optimizations on AMD GPUs and how 800M+ tensors
Hands-On Workshop
- The afternoon workshop provided attendees hands-on experience through deploying vLLM on AMD GPUs, with four AMD GPU servers provided for live experimentation. Participants in the lab demo ran some of the latest GPUs, with home teams ready to give deployment feedback.
Evening Sessions: Lightning Talks & Networking
- Performance AI - Yizhou, an LLCPP+ model and AudioLLM for Southeast Asia
- AMD Instinct™ MI300X - Accelerating LLM inference with AMD ROCm™ and vLLM
- Multimodal Inference - Carefully fine-tuned vLLM implementation enabling encoders for multimodal models The evening ended with a lively social mixer—a chance for participants to chat, connect, and spark collaborations.
👥 Speakers
The event featured an incredible lineup of contributors and industry experts, including:
- Chen Zhang (TsinghuA, vLLM Committer)
- Xiyou Liang (HKUST, vLLM Maintainer)
- George Sung, Bruce Ma, Haichen Zhang (AMD AI Software & Product teams)
- Tim Jian Tan (Principal Engineer, EmbeddedLLM & AMD optimization expert)
- Dr. Bin Shao (Principal Founder & CTO, Embedded LLM)
- Dr. Yat Hei Cheung (Principal LLM Solutions Architect, Embedded LLM)
- Yingyu He (APTAR, PR)
- Camille Du (SEA AI Lab)
The Inaugural vLLM Asia Developer Day marked a milestone in building an open, collaborative ecosystem for LLM inference innovation in Asia. By combining cutting-edge research, practical engineering know-how, and community-driven collaboration, this event set the stage for a stronger, more connected AI developer community across the region.
We thank all 100+ participants, speakers, and partners for making this event such a success. The energy, curiosity, and technical depth from the community inspire us to keep building—and this is just the beginning!
