Han Li
Senior Staff Machine Learning Researcher, Shopify
I am a machine learning researcher at Shopify. My work focuses on agentic foundation models, vision-language modeling, and user simulation for AI agents. Before joining Shopify, I was an Applied Scientist at Amazon Stores Foundational AI, where I led efforts in large language model pretraining, as well as vision-language model pretraining and post-training.
I obtained my Ph.D. in Computer Sciences from the University of Wisconsin-Madison, advised by Prof. AnHai Doan, and received my bachelor's degree in Computer Science from Tsinghua University.
Updates
- [NEW] ShopGym accepted to the ICML 2026 RLxF Workshop.
- [NEW] We open-sourced ShopGym, an integrated framework for realistic simulation and scalable benchmarking of e-commerce web agents. ShopGym provides self-contained sandbox shops and grounded evaluation tasks, making agent evaluation more reproducible, inspectable, and scalable.
- Published an engineering deep dive on fine-tuning a tool-calling agent that converts plain-English merchant requests into executable Shopify Flow automations.
- One paper was accepted to ACL 2026.
- Published an engineering deep dive on SimGym, a synthetic-shopper platform for running realistic storefront A/B tests at scale before touching live traffic.
Selected Publications
-
SimGym: A Framework for A/B Test Simulation in E-Commerce with Traffic-Grounded VLM Agents
-
SimPersona: Learning Discrete Buyer Personas from Raw Clickstreams for Grounded E-Commerce Agents
-
LongLeader: A Comprehensive Leaderboard for Large Language Models in Long-context Scenarios
-
Multimodal Instruction Tuning with Hybrid State Space Models
-
Deep Learning for Blocking in Entity Matching: A Design Space Exploration
-
Continuous Learning for Large-Scale Personalized Domain Classification
-
Deep Learning for Entity Matching: A Design Space Exploration