Qi Chen
I am an Applied Scientist at AWS AI Labs, where I build LLM-powered coding agents. My research focuses on test-time scaling, benchmark design, and agentic reinforcement learning, with the goal of scaling agents toward fully autonomous, long-horizon software development tasks. I am particularly interested in coding as a universal interface – a medium through which humans and agents can collaboratively shape the digital world.
Prior to Amazon, I received my Ph.D. in Applied Mathematics from Peking University (2023), advised by Prof. Zhouchen Lin, Prof. Yisen Wang, and Prof. Jiansheng Yang. My doctoral research focused on deep equilibrium models (infinitely deep neural networks) for context learning, with applications to computer vision and graph learning.
I’m open to collaboration! Feel free to email me if you’re interested in working together.
news
| Dec 2025 | Kiro autonomous agent was launched and showcased in Matt Garman’s keynote at AWS re:Invent 2025; I led the end-to-end agent evaluation and benchmark development. |
|---|---|
| Dec 2024 | Led the initialization and launch of the /doc feature for Amazon Q Developer, the first repository-level code documentation generation system, enabling analysis of codebases far exceeding LLM’s context windows; showcased in Matt Garman’s keynote at AWS re:Invent 2024. |
| May 2024 | Joined AWS AI Labs as an Applied Scientist, working on LLM for coding. |
| Feb 2024 | CREAD, a classification-restoration framework for long-tail user watch-time prediction, was accepted at AAAI 2024 and deployed at scale in the Kwai App. |
| Jul 2023 | Earned a Ph.D. in Applied Mathematics from Peking University, with a dissertation on deep equilibrium models for context modeling. |
selected publications
- AAAICREAD: A Classification-Restoration Framework with Error Adaptive Discretization for Watch Time Prediction in Video Recommender SystemsIn AAAI, 2024
- TIPEquilibrium Image Denoising with Implicit DifferentiationIEEE Transactions on Image Processing, 2023
- BigDataEfficient and Scalable Implicit Graph Neural Networks with Virtual EquilibriumIn IEEE International Conference on Big Data (Big Data), 2022