Qitian Wu

About

I am a Research Scientist at Google Research, working on LLM post-training and agents, building hands-on across the stack from data and training environments to evaluation and production deployment. My recent projects:

• Gemini post-training for Text-to-SQL generation: post-trained Gemini to the top of the BIRD Text-to-SQL leaderboard. (Synthetic data generation, sandbox environments with tool use, reward models)
• Agents for systems optimization: built a closed-loop agent that optimizes production code from evolving benchmark feedback, deployed in compiler systems. (Agentic harness, long-horizon reasoning, endpoint serving)
• Gemini reasoning for network-state observability: partnered with infrastructure teams to scope large-scale cluster telemetry into an LLM reasoning task. (Data curation and filtering, evaluation pipelines)

Previously, as an Eric and Wendy Schmidt Center Postdoctoral Fellow at the Broad Institute of MIT and Harvard (with Caroline Uhler), I led the pretraining of a multi-modal foundation model from scratch, from data pipelines to distributed training on GPU clusters, and applied it to discovery of therapeutic targets in biomedicine (paper).

During my PhD, I worked as a part-time research intern at Amazon Web Services for three years, designing linear-complexity attention that scales Transformers to 1M-token contexts on a single 16GB GPU, and releasing the resulting models as widely used open-source projects (NodeFormer, SGFormer, DIFFormer). Earlier, I was an applied research intern at Tencent WeChat, building deep learning models for industry-scale recommender systems.

I received my Ph.D., M.E. and B.E. degrees from Shanghai Jiao Tong University, with my PhD advised by Junchi Yan. My research has been published at top venues including NeurIPS, ICML, ICLR, JMLR and TPAMI, and has been recognized with the Microsoft Research PhD Fellowship, the Baidu PhD Fellowship and the Elsevier/Stanford World's Top 2% Scientists list. I serve as an Area Chair for NeurIPS and ICLR.

Selected Publications

Full list on Google Scholar.

LLMs, Agents & Foundation Models

Scalable Transformer Architectures

Theory of Machine Learning & Generative Models

Generalization under Distribution Shift

Selected Awards

Leadership & Service

Conferences	Area Chair, NeurIPS and ICLR; Senior Program Committee, AAAI and IJCAI
Journals	Action Editor, TMLR; Editorial Board Reviewer, JMLR