Hengyue Liang (梁恒岳)

Applied Scientist, Applied AI, Amazon

Bellevue, WA, USA

Email:

lianghengyue1993 (at) outlook (dot) com

Links:

| Google Scholar | Resume | LinkedIn | Github |

My name is Hengyue. I received my Ph.D. in Electrical and Computer Engineering from the University of Minnesota, Twin Cities (May 2025), advised by Prof. Ju Sun. I am currently an Applied Scientist at Amazon (Applied AI - World-Wide Store), based in Bellevue, WA.

My work focuses on building and shipping production Agentic AI systems at enterprise scale — browser-operating web agents, the model interaction layer behind them, inference and cost optimization, agent runtimes, and the evaluation infrastructure that keeps them reliable. Recent projects include a browser-operating LLM agent deployed in production, an LLM-as-judge reliability system for monitoring live agent traffic, and an autonomous pipeline that debugs and repairs deployed agents. See Work Experience for details.

I keep an active research agenda on LLM agent architectures and agent evaluation (KDD 2025 workshop). My earlier academic work spans trustworthy AI, adversarial robustness, selective classification, computer vision, and deep RL for robotic manipulation — 10 peer-reviewed papers in TMLR, CVPR, ICRA and RA-L, with ~500 citations (h-index 8). I also review for NeurIPS, CVPR, ICRA, IROS and TMLR.

Before UMN, I earned an M.S. in Electrical Engineering from Chalmers University of Technology (Sweden) and a B.Eng. from Shanghai Jiao Tong University (China).