News

  • [2026.05] Invited by Prof. Roberto V. Zicari to co-author a project on Trustworthy Generative AI Development and Use for ODBMS.org.
  • [2026.03] Prompt Bank was awarded the 2025 DeepMind Data Science Breakthrough Award.
  • [2025.07] Gemini 2.5: Core Contributor for Multimodal Long Context PDF understanding and first Google I/O contribution!
  • [2025.07] Presented LLM Loss Analyzer demo at Google DataVerse Summit 2025.
  • [2025.04] Presented Autorater Noise Predictor demo at Google Data Science Summit 2025.
  • [2024.06] Grail went public on the NASDAQ.
  • [2024.05] Graduated as a Fusion Fellow for the 2024 Venture Fellowship Program.
  • [2020.04] The Wall Street Journal reported our work on the Stanford Hospital Resource Calculator during the pandemic.
  • [2019.10] Won First Place at the 2019 Stanford Citadel Datathon, tackling real-world social challenges.

Experiences

Google, Core ML Data Jan 2025 - Present
Research Data Scientist | Mountain View, CA
• Gemini Post-training: Improved Gemini's Flash performance on financial document Q&A from 54.42% (Gemini 2.0) to 66.50% (Gemini 2.5). The Pro performance has improved from 72.73% (Gemini 2.5) to 81.55% (Gemini 3.0).
• Frontier Safety: Conducted fine-tuning and APO experiments, achieving 0.969 recall for misuse detection pipelines. Developed an agentic approach for long-term user trajectory analysis for cyber attack detection.
• Automated Tooling: Architected an end-to-end Automated Loss Pattern Analysis Tool used across Google teams, onboarded factuality and reasoning eval to Prompt Bank, and developed an agentic approach for benchmark generation for post train eval refactor.
Grail, LLC Dec 2022 - Oct 2024
Machine Learning Scientist | Menlo Park, CA
• Cancer Detection: Focused on multi-modal data fusion and classifier calibration for early signal predictors.
Uber Technologies Nov 2021 - Dec 2022
Data Scientist, Driver Pricing | San Francisco, CA
• Algorithmic Pricing: Team led marketplace-wide launch of driver pricing models to 75% of Uber's global business, with over 4% user growth in South America market.

Education

Stanford University Sep 2016 - Dec 2021
Ph.D. in Computational Mathematics | GPA: 3.97/4.3
Advisors: Prof. Peter Glynn, Prof. Jose Blanchet.

Teaching Experiences:
• MS&E 325: Optimal Transport in Operations Research, Statistics, and Economics
• MS&E 324: Stochastic Methods in Engineering
• MS&E 121: Introduction to Stochastic Modeling
Stanford Graduate School of Business Jul 2021
Stanford Ignite Executive Program
Certificate in Innovation and Entrepreneurship. Won the top 3 pitch in the final pitch competition.
Peking University Sep 2012 - Jul 2016
B.S. in Applied Mathematics | GPA: 3.78/4.0
Double Major in Economics.
Advisors: Prof. Lin Lin (UC Berkeley) and Prof. Huijun Fan (Peking University).
Academic Tree

Selected Publications

Gemini 2.5 Figure
Gemini 2.5: Pushing the frontier with advanced reasoning, multimodality, long context, and next generation agentic capabilities Google Gemini team | arXiv, 2025
Optimal Transport Figure
Surgical Scheduling via Optimal Transport Jin Xie, et al. | Submitted to Operations Research, 2024+
Distributionally Robust Optimization Figure
Asymptotic Analysis of the Profile Function in Distributionally Robust Optimization Jin Xie, et al. | Submitted to AAAI, 2023+
Simulation Figure
Simulation: address the limitation of consideration of non-controllable part of scheduling Jin Xie, et al. | Healthcare Management Science, 2023
Ph.D. Thesis Figure
Optimal transport and healthcare operations Jin Xie | Ph.D. Thesis, Stanford University, 2022
COVID-19 Model Figure 1
Interactive Model to Estimate Bed Demand for COVID-19-Related Hospitalization Developed by Stanford Medicine-Engineering Partnership Jin Xie, et al. | Health Management, Policy and Innovation, 2020
COVID-19 Model Figure 2
A model to estimate bed demand for COVID-19 related hospitalization Jin Xie, et al. | medRxiv preprint, 2020

Honors & Awards

Invited Talks

Academic Services and Volunteer

Leadership & Community Service
• Vice President of Business Development, Stanford Chinese Entrepreneur Organization
• President, SIAM (Society for Industrial and Applied Mathematics) Student Chapter, Stanford
• Board Member, Women in Computational Sciences and Engineering, Stanford
• Vice President, Stanford Oriental Zither & Qin Association
• Vice President, Teenage Volunteers Association, School of Mathematical Sciences, Peking University
Journal & Conference Reviewer
Management Science, Mathematical Programming, Annual Conference on Neural Information Processing Systems (NeurIPS), Artificial Intelligence and Statistics (AISTATS), IEEE International Conference on Multimedia Information Processing and Retrieval (MIPR), IEEE Conference on Decision and Control (CDC), The Design Theory and Methodology (DTM).

Miscellaneous

Outside of work I love art, sports, and travelling.

  • I write in career mentorship and AI research across multiple platforms (30K+ followers).
  • I love skiing, trekking, jazz, and pickleball. I’ve run the SF Half Marathon twice, and to celebrate graduating from college, I walked one-third of South Korea’s coastline!
  • Semi-professional Guzheng performer (amateur level 10). Outstanding award in the Stanford Baipu International Chinese music competition (Cover) and annual concert performances at the Stanford Department of Music with the Baipu Guzheng Ensemble.
  • Amateur actress/producer with the Feiyu Theater. Main productions: theater performance for The Ride Down Mt. Morgan by Arthur Miller in 2020 and theater producer for The Message by Mai Jia in 2024.