ChangyiYang's Site

Reduce MHA to its arithmetic intensity and you get exactly 1.

TL;DR: Agent execution comes down to one question — how do you sensibly parallelize a single task across multiple agents?

TL;DR In LLM RL post-training, every optimizer step on the trainer has to push fresh weights into the inference engine doing rollouts.

I recently started looking into LoRA + RL, and as a side project I dug into how SGLang implements LoRA serving.

When you spin up an LLM with SGLang, you get this satisfying sequence: progress bars flying, safetensors loading, and then — boom — the model starts generating tokens.

Ref: https://www.bilibili.com/video/BV1darmBcE4A/ I listened to the n+e podcast and wrote down the key points and my takeaways.

Ref: Ilya Sutskever’s talk on AI development and future directions I recently listened to a fascinating talk by Ilya Sutskever, co-founder of OpenAI, discussing the current state and future directions of AI development.

Ref: https://www.bilibili.com/video/BV1s6pgzLE3y/?

Ref: https://www.xiaoyuzhoufm.com/episode/688cc1cc8e06fe8de7d920cd?

References: Interview on AI Infrastructure Career Path and Insights from AWS/Meta Job Transition and Interview Preparation for AI Infra in North America AI infrastructure has emerged as a key enabler in the acceleration of artificial intelligence technolo…

From AIEngineer: RAG Agents in Prod: 10 Lessons We Learned In the rush toward AI adoption, everyone is chasing cutting-edge models, fine-tuning pipelines, and building impressive prototypes.

References: YouTube Video by Jordan Has No Life Kafka Summary Notes This article summarizes my personal understanding after watching the aforementioned video and reading the Kafka paper.

The idea of building a personal site started back in September last year.