Wenbo Chen

Applied Scientist at Amazon

I build AI systems that make decisions — from autonomous agents that plan and act, to data-driven methods that enable real-time operations at scale. My work bridges AI agents, LLMs, and data-driven optimization.

Currently an Applied Scientist at Amazon working on seller-facing AI agents. ML PhD from Georgia Tech / NSF AI4OPT with Pascal Van Hentenryck. Our SkillsBench is increasingly adopted as an agentic evaluation by frontier labs (Qwen, HY3), and my work on scalable data-driven decision making has seen major industrial adoption.

Email Scholar GitHub Twitter

Research

My research focuses on enabling AI systems to make complex decisions autonomously and reliably. I am particularly interested in how LLM-based agents can acquire, compose, and execute procedural skills to solve real-world tasks — and how we can verify and trust their decisions.

AI Agents LLM Skills & Evaluation Decision Making Trustworthy AI Data-Driven Optimization

News

May 2026 Co-organizing the First Workshop on Agent Skills at ACM CAIS 2026, San Jose, CA.
Apr 2026 Paper on LLM hallucination detection accepted at ACL 2026 (Main Conference).
Apr 2026 Paper on RL with combinatorial actions accepted at ICML 2026 (Spotlight).
Feb 2026 SkillsBench released — benchmarking agent skills across diverse tasks.
Jul 2025 "Outbound Load Planning in Parcel Delivery Service Networks" accepted in INFORMS Transportation Science.
Oct 2025 "Boosting Column Generation with GNNs for Joint Rider Trip Planning and Crew Shift Scheduling" published in Transportation Research Part E.
Jun 2024 Joined Amazon as an Applied Scientist, working on LLM Agents and Trustworthy AI.
May 2024 "Compact Optimality Verification for Optimization Proxies" accepted at ICML 2024.
Apr 2024 Defended my PhD thesis! News.
Mar 2024 Research highlighted in AI Magazine (NSF's National AI Institutes).
Feb 2024 Invited talk at Rice University CMOR Colloquium Series.
Jan 2024 Won VNN-COMP'23 Award for outstanding benchmark.
Nov 2023 Awarded Anderson-Interface Fellowship for Excellence in Research (Energy & Sustainable Systems).
Sep 2023 Two papers accepted in IEEE Transactions on Power Systems. Paper and Paper.
Jul 2023 One paper accepted in IEEE Transactions on Power Systems. Paper.
May 2023 Won thesis pitch competition (ML & manufacturing track) at IISE Annual Meeting 2023!
May 2023 Invited presentation: End-to-End Learning and Optimization thrust for AI4OPT External Advisory Board.
Apr 2023 Invited presentation: "End-to-End Feasible Optimization Proxies for Large-Scale Economic Dispatch" at INFORMS Annual Meeting 2023.
Mar 2023 Invited presentation: "Confidence-Aware GNNs for Learning Reliability Assessment Commitments" at IISE Annual Meeting 2023.
Mar 2023 Featured in NSF AI4OPT newsletter student highlights.
Jan 2023 Attended 2023 Grid Science Winter School and Conference.
Oct 2022 Invited presentation: "Learning Optimization Proxies for Large-Scale Security-Constrained Economic Dispatch" at INFORMS Annual Meeting 2022.

Selected Publications

Rethinking Evaluation for LLM Hallucination Detection

W. Chen, V. Padmanabhan, T. Giyahchi, E. Wong, L. Akoglu

ACL 2026

Paper
Latent Spherical Flow Policy for RL with Combinatorial Actions

L. Kong, A. Satish, H. Jiang, A. Kangaslahti, A. Ma, W. Chen, et al.

ICML 2026 (Spotlight)

Paper

Spotlight
SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse Tasks

X. Li, W. Chen, Y. Liu, et al.

Preprint, 2026

Paper
Compact Optimality Verification for Optimization Proxies

W. Chen, H. Zhao, M. Tanneau, P. Van Hentenryck

ICML 2024

Paper
End-to-End Feasible Optimization Proxies for Large-Scale Economic Dispatch

W. Chen, M. Tanneau, P. Van Hentenryck

IEEE Transactions on Power Systems, 2023

Paper

View all publications →