Hi, I'm Tesi Xiao.
Senior Applied Scientist at Amazon Search, building machine learning systems that serve hundreds of millions of customers — from ideation to production.
I received my Ph.D. in Statistics from UC Davis, advised by Prof. Krishna Balasubramanian. My work bridges optimization theory and large-scale ranking systems, with a focus on principled methods that actually ship.
Research Interests
- Learning to Rank
- Multi-Objective Optimization
- LLMs for Information Retrieval
- Online Decision Making
- Stochastic Optimization
- Robust Deep Learning
News
- Paper on test-time control for LLM reasoning accepted to ICML 2026.
- Paper on multi-objective DPO accepted to UAI 2025.
- First HCI paper on multi-objective ranker evaluation accepted to IUI 2025.
- Stochastic bilevel optimization work published in JMLR.
- Papers at ICLR 2024 and AISTATS 2024.
Experience
Senior Applied Scientist · Amazon, Core Search
Applied Scientist II · Amazon, Search Science and AI
Applied Scientist Intern · Amazon, Search Science and AI
Research Scientist Intern · ByteDance, Applied Machine Learning
Selected Papers
Optimal Algorithms for Stochastic Bilevel Optimization under Relaxed Smoothness Conditions
JMLR
A Projection-free Algorithm for Constrained Stochastic Multi-level Composition Optimization
NeurIPS 2022
How Does Noise Help Robustness? Explanation and Exploration under the Neural SDE Framework
CVPR 2020
Service
- Top Reviewer: NeurIPS 2023, UAI 2023
- Reviewer: NeurIPS, ICML, ICLR, UAI, AISTATS, COLT, SIAM Journal on Optimization
- Talks: Amazon Shopping Science Summit, Amazon ML Conference, SIGIR-AP, INFORMS