Kunyu Shi

I lead an LLM post-training team at Alibaba (2024–present), with groups in Hangzhou and Sunnyvale. We build agentic foundational models through continual pre-training, supervised fine-tuning, reinforcement learning. Our models power AI services on Alibaba.com and Accio.com. We are hiring!

Previously, I was a Senior Applied Scientist at AWS AI Labs (2020–2024), where I worked in Stefano Soatto’s group at Caltech on document understanding and large foundation models (LLM and VLM). I contributed to several AWS AI services including Amazon Textract and Amazon Bedrock.

I received my M.S. in Computer Science from the University of California, San Diego (UCSD) in 2020 and my B.Eng. in Electrical Engineering from the University of Electronic Science and Technology of China (UESTC) in 2018. Prior to industry, I was a research assistant at the Machine Learning, Perception, and Cognition Lab (mlPC) at UCSD with Prof. Zhuowen Tu (2019–2020) and at the Big Data Research Center at UESTC with Prof. Tao Zhou (2016). I also interned at McMaster University with Prof. Paul W. Ayers (2017).

My academic and industrial research interest includes detection, segmentation, large scale multi-modality learning, large language model and document understanding.