Experience

  1. Senior Researcher

    Microsoft Research
    • Machine Intelligence team, working on efficient AI.
    • Leading research on learning dynamics and optimization for foundation models.
    • Project Next-Gen Optimizer: Developing theoretical and practical frameworks to accelerate training while reducing resource usage.
    • Project Causica: Lead developing temporal causal models for time-series data across discrete and continuous-time settings.
    • Collaborate with Global Channel Sales (GCS) to apply causal ML for revenue growth.
  2. Researcher

    Microsoft Research
    • Project Causica: Lead causal time-series projects providing end-to-end causal inference pipelines for time-series data.
    • Led collaboration with Eedi on applying causality to education.
    • Lead organizer for NeurIPS 2022 CausalML competition; co-organized NeurIPS 2022 score-based method workshop.
  3. Research Intern

    Microsoft Research
    • Mentor: Dr. Cheng Zhang
    • Developed a deep generative model and an inference method for active learning and prediction with accurate uncertainty quantification in small data regime. Published in NeurIPS 2019 with two patents.

Education

  1. PhD Engineering (Machine Learning)

    University of Cambridge
    Machine Learning Group. Supervisor: Dr. Jose Miguel Hernandez-Lobato. Thesis: Advances in approximate inference: combining VI and MCMC and improving on Stein discrepancy. Research focus: approximate inference, generative model, MCMC sampling, Bayesian learning.
    Read Thesis
  2. MPhil Machine Learning, Speech and Language Technology

    University of Cambridge
    Project: Wasserstein Generative Adversarial Network supervised by Prof. Richard E. Turner. Co-supervised by Dr. Yingzhen Li and Dr. Mark Rowland.
  3. MEng Information and Computer Engineering

    University of Cambridge
    Honours pass with distinction in exams and project. Project: Sampling method for Indian buffet process. Supervisor: Prof. Zoubin Ghahramani. Co-supervised by Dr. Ge Hong and Dr. Nilesh Tripuraneni.
  4. BA Electrical and Information Science

    University of Cambridge
    First class in all three years. Top 1 in Gonville and Caius College.
Skills & Hobbies
Research Areas
Learning dynamics and optimization
Foundation models and LLMs
Causality
Approximate inference
Deep generative models
Technical Skills
Python
MatLab
Hobbies
GYM
Cooking
Music
Awards
ICLR Travel Award
ICLR
Travel award (USD 1,000).
January 2019
Student Scholarship
Machine Learning Summer School
Student scholarship (EUR 1,200).
January 2018
Cambridge Trust CSC Scholarship
University of Cambridge
Tuition fee plus GBP 14,400 per year.
January 2017
Senior Scholarship (2013-2015)
University of Cambridge
January 2015
Languages
70%
English
100%
Chinese (Native)