Hansi Zeng — UMass Amherst

About

I am a final‑year PhD student in Computer Science at the University of Massachusetts Amherst, advised by Prof. Hamed Zamani.

Previously, I was a research intern at Google DeepMind, Amazon Science, and Snapchat.

My research lies at the intersection of information retrieval and language model agents. My goal is to bridge complex user questions and accurate, grounded answers whose supporting evidence may be buried across web‑scale corpora. To that end, I develop scalable and robust retrieval systems through sparse, dense, and generative retrieval modeling, design novel learning‑to‑rank algorithms, and study their scaling behavior. I also study how to train language model agents with on‑policy reinforcement learning to use retrieval and other tools, acquire information, reason over evidence, and solve complex questions.

I am currently on the job market! I am actively seeking full-time industry roles (e.g., Research Scientist) starting in Summer or Fall 2026.

Experience

	Snapchat · Part-time Research Intern Sep 2025 – Present
	Google DeepMind · Student Research Intern May 2024 – Mar 2025
	Amazon · Applied Scientist Intern May – Dec 2023 · May – Aug 2025

Selected Publications (Full list on Google Scholar)

CoSearch: Joint Training of Reasoning and Document Ranking via Reinforcement Learning for Agentic Search

Hansi Zeng, Liam Collins, Bhuvesh Kumar, Neil Shah, Hamed Zamani

In Submission
LLM Development at Fixed Scale: Lightweight Proxies to Overcome the Perplexity Illusion

Hansi Zeng, Kai Hui, Honglei Zhuang, Zhen Qin, Zhenrui Yue, Hamed Zamani, Dana Alon

In Submission
SynPlanResearch-R1: Encouraging Tool Exploration for Deep Research with Synthetic Plans

Hansi Zeng, Zoey Li, Yifan Gao, Chenwei Zhang, Xiaoman Pan, Tao Yang, Fengran Mo, Jiacheng Lin, Xian Li, Jingbo Shang

arXiv 2026

[Paper] [Code]
Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning

Bowen Jin, Hansi Zeng, Zhenrui Yue, Jinsung Yoon, Sercan Arik, Dong Wang, Hamed Zamani, Jiawei Han

COLM 2025

[Paper] [Code]
Scaling Sparse and Dense Retrieval in Decoder-Only LLMs

Hansi Zeng, Julian Killingback, Hamed Zamani

SIGIR 2025

[Paper] [Code]
Hypencoder: Hypernetworks for Information Retrieval

Julian Killingback, Hansi Zeng, Hamed Zamani

SIGIR 2025

[Paper] [Code]
Inference Scaling for Long-Context Retrieval Augmented Generation

Zhenrui Yue, Honglei Zhuang, Aijun Bai, Kai Hui, Rolf Jagerman, Hansi Zeng, Zhen Qin, Dong Wang, Xuanhui Wang, Michael Bendersky

ICLR 2025 (Oral)

[Paper]
Planning Ahead in Generative Retrieval: Guiding Autoregressive Generation through Simultaneous Decoding

Hansi Zeng, Chen Luo, Hamed Zamani

SIGIR 2024

[Paper] [Code]
Scalable and Effective Generative Information Retrieval

Hansi Zeng, Chen Luo, Bowen Jin, Sheikh Muhammad Sarwar, Tianxin Wei, Hamed Zamani

WWW 2024 (Oral)

[Paper] [Code]
Language Models as Semantic Indexers

Bowen Jin, Hansi Zeng, Guoyin Wang, Xiusi Chen, Tianxin Wei, Ruirui Li, et al.

ICML 2024

[Paper] [Code]
Soft Prompt Decoding for Multilingual Dense Retrieval

Zhiqi Huang, Hansi Zeng, Hamed Zamani, James Allen

SIGIR 2023

[Paper]
A Personalized Dense Retrieval Framework for Unified Information Access

Hansi Zeng, Surya Kallumadi, Zaid Alibadi, Rodrigo Nogueira, Hamed Zamani

SIGIR 2023

[Paper]
Curriculum Learning for Dense Retrieval Distillation

Hansi Zeng, Hamed Zamani, Vishwa Vinay

SIGIR 2022 · Best Short Paper Award

[Paper] [Code]
A Zero Attentive Relevance Matching Network for Review Modeling in Recommender System

Hansi Zeng, Zhichao Xu, Qingyao Ai

ECIR 2021

[Paper]

Services

Reviewer: SIGIR, WWW, COLM, ICML, NeurIPS, TOIS, CIKM, ACL, ECIR