PeopleofSuperintelligence

Researcher

Xin Wang

Xin Wang is a senior researcher at Microsoft Research in Redmond. She completed a bachelor's degree in computer science at Shanghai Jiao Tong University in 2015 and a Ph.D. at the University of California, Berkeley in 2020, working with Trevor Darrell and Joseph Gonzalez. Her 2020 papers included BDD100K, a driving dataset of 100K videos and 10 tasks, and Frustratingly Simple Few-Shot Object Detection. She started at Microsoft Research in May 2021. In 2023 and 2024 she coauthored Gorilla, a model that writes API calls, Textbooks Are All You Need, which introduces the phi-1 code model, and the phi-4 technical report on a 14-billion-parameter language model.

Works
1

Publication numbers from OpenAlex, as of 17 October 2026. Citations and h-index are hidden because this OpenAlex record may be incomplete or mix in other authors.

Timeline

  1. 2011–2015

    A bachelor's degree in Shanghai

    From September 2011 to June 2015 she completes a bachelor's degree in computer science in the IEEE Pilot Class at Shanghai Jiao Tong University.

  2. 2015–2020

    A doctorate at Berkeley

    From August 2015 to December 2020 she completes a Ph.D. in computer science at the University of California, Berkeley, working with Trevor Darrell and Joseph Gonzalez in the BAIR Lab and the RISE Lab. In November 2020, UC Berkeley selects her as one of its Rising Stars in EECS.

  3. 2020

    Driving video and few-shot detection

    BDD100K, presented at CVPR 2020 with Fisher Yu, Trevor Darrell, and other coauthors, is a driving dataset of 100K videos and 10 tasks. At ICML 2020, Frustratingly Simple Few-Shot Object Detection shows that fine-tuning only the last layer of existing detectors on rare classes outperforms meta-learning methods.

  4. May 2023

    Gorilla for API calls

    In May 2023 she introduces Gorilla with Shishir Patil, Tianjun Zhang, and Joseph Gonzalez. Gorilla is a finetuned LLaMA-based model that surpasses the performance of GPT-4 on writing API calls.

  5. June 2023

    Textbooks are all you need

    In June 2023 she coauthors Textbooks Are All You Need, which introduces phi-1, a Transformer-based model with 1.3B parameters for code. Training uses textbook-quality data from the web and synthetic textbooks and exercises generated with GPT-3.5.

  6. December 2024

    The phi-4 language model

    In December 2024 she coauthors the phi-4 technical report. It presents a 14-billion-parameter language model developed with a training recipe focused on data quality, and the training incorporates synthetic data throughout.

  7. Since 2021

    Senior researcher in Redmond

    She is a senior researcher at Microsoft Research in Redmond. The appointment begins in May 2021. She joins the Physics of AGI group led by Sébastien Bubeck and stays associated with the Computer Vision Group.

Sources and credits

Photo: We have no published photo of Xin Wang yet. Send us a free-licensed photo.

  1. Xin Wang's Personal Homepage xinw.ai. Checked 16 October 2026.
  2. Xin Wang - People @ EECS at UC Berkeley xinw.ai. Checked 16 October 2026.
  3. [1805.04687] BDD100K: A Diverse Driving Dataset for Heterogeneous Multitask Learning arxiv.org. Checked 16 October 2026.
  4. [2003.06957] Frustratingly Simple Few-Shot Object Detection arxiv.org. Checked 16 October 2026.
  5. [2305.15334] Gorilla: Large Language Model Connected with Massive APIs arxiv.org. Checked 16 October 2026.
  6. Gorilla gorilla.cs.berkeley.edu. Checked 16 October 2026.
  7. Textbooks Are All You Need - Microsoft Research microsoft.com. Checked 16 October 2026.
  8. [2306.11644] Textbooks Are All You Need arxiv.org. Checked 16 October 2026.
  9. Phi-4 Technical Report - Microsoft Research microsoft.com. Checked 16 October 2026.
  10. [2412.08905] Phi-4 Technical Report arxiv.org. Checked 16 October 2026.