Breaking the AI Inference Bottleneck with AWS Trainium and vLLM
The surge in reasoning models and agentic AI has exposed a critical infrastructure gap: conventional deployment approaches can't deliver the performance these applications demand without prohibitive costs. Learn how AWS Trainium and vLLM solve this challenge, enabling production-scale AI that's both lightning-fast and cost-effective.
Join us for an engaging evening featuring a hands-on workshop, tech talks, networking opportunities, and delicious food and drinks. Seats are limited. We encourage you to register in advance to secure your spot.
What to expect:
2:30 pm - Welcome and Registration
3:00 pm - AWS Overview Introduction
3:20 pm - Hands-on Workshop featuring AWS Trainium and vLLM
4:00 pm - Tech Talk: Optimizing Performance for Your Applications
4:20 pm - Tech Talk: Disaggregated Prefill on AWS Trainium and vLLM for Enhanced Performance
4:40 pm - Tech Talk: vLLM's Trainium Optimization Roadmap: Performance Enhancements and Features
5:00 pm - Networking and Happy Hour
Speakers:
Jim Burtoft, Senior Solutions Architect, AWS
Pinak Panigrahi, Senior GenAI Architect, Annapurna ML, AWS
Mrinal Shukla, Serving Acceleration Senior Manager, Annapurna ML, AWS
Note
Please bring a physical government-issued photo ID. Digital IDs won't be accepted.
There is no scooter/bike parking available in the building. If you plan to bring one, you will need to find parking options nearby.
All attendees must be 18 or older to attend this event at the space.