Emily Webber, Principal Machine Learning Specialist Solutions Architect at Amazon Web Services (AWS), has taken to LinkedIn to outline plans for a Trainium-focused workshop scheduled at NeurIPS 2025.
AWS develops cloud and machine learning infrastructure, including Trainium, its custom accelerator designed to improve the speed and cost-efficiency of training large models across distributed systems.
NeurIPS, the Conference on Neural Information Processing Systems, is one of the largest global events for machine learning and AI research, bringing together researchers, engineers, and developers working on cutting-edge advances in the field.
The workshop, titled Checkmate: Fine-Tune Your Own Small Language Model for Real-Time Chess Gameplay, will take place on December 2 at the San Diego Convention Center.
Workshop uses chess to test model reasoning and explanation
In her post, Webber wrote that she is “thrilled to announce a hands-on workshop about playing chess on AWS Trainium,” noting that the session will explore how small language models handle rule-based reasoning and strategy. Webber said: “Chess is a great learning environment for many aspects in machine learning. The game itself is well handled by CPU-based gaming engines, but the ability of LLMs to understand these discrete rules and common strategies is vastly under-explored.”
She added that combining gameplay with dialogue creates space to examine how models reason, explain moves, and engage in back-and-forth interaction.
Attendees will fine-tune models and compete on a live leaderboard
Participants will access an AWS environment via QR code and fine-tune a Qwen-based language model to play chess in real time. The workshop includes inference demonstrations and a leaderboard hosted by AIcrowd and AGI House, allowing attendees to submit their models and monitor performance during the session.
AWS will also demonstrate how Trainium accelerates both training and inference workloads. The chip is built to deliver higher throughput than traditional GPU-based systems for many training tasks, particularly when running repeated fine-tuning cycles on compact or mid-sized LLMs.
Webber described the event as an opportunity to build skills in a controlled environment, saying it gives participants “the ideal environment to grow your skills, both as a data scientist and as a chess enthusiast.”
Webber confirmed that she will lead the workshop with colleagues from AWS AI/ML and partners from AIcrowd and AGI House. She wrote: “We’re bringing all of this to you and more at NeurIPS, and I couldn’t be happier to team up with Rocky Yu at AGI House, Sharada Prasanna Mohanty at AI Crowd, Henry Yin, and the stellar builders at AWS with Jim Burtoft Niithiyn Vijeaswaran Megan McKenzie Josh Longenecker Ruchi Bhatia Armin A. and so many more.”
Registration is open for NeurIPS attendees, with limited spaces available. Webber added that a version of the session will also be offered at AWS re:Invent in December.
The ETIH Innovation Awards 2026
link
About the Author
