.. ..
Berkeley,CA
Overview
2
2
years of professional experience
Work History
Research Assistant
Berkeley AI Research
Berkeley, California
05.2025 - Current
- Mentored by PhD student Yiheng Du and Professor Aditi Krishnapriyan.
- Built and fine-tuned machine learning models to predict complex physical turbulence systems.
- Applied different loss functions (data-driven, gradient-based, physics-informed) to improve model accuracy and energy conservation stability.
- Ran large-scale experiments on supercomputing clusters, managing datasets, logs, and checkpoints.
- Analyzed results using error metrics and visualizations to evaluate model performance over time.
Research Assistant
University of Science and Technology of China
Beijing, China
07.2023 - 08.2025
- Mentored by Professor David P. Woodruff from Carnegie Mellon University
- Evaluated the Confidence-based Group Label Assignment (CGL) method, uncovering significant limitations and providing empirical evidence of its insufficiencies in training fair classifiers.
- Proposed and developed novel modifications to the CGL method to enhance its ability to effectively train fair classifiers, improving the robustness and fairness of the model.
- Authored and published a research manuscript titled "In-Depth Exploration and Potential Improvements on Learning Fair Classifiers with Partially Annotated Datasets" in the proceedings of CONF-SEML 2024 (Print ISSN: 2755-2721).
Teaching Intern
BASIS International
Guangzhou, China
08.2023 - 06.2024
- Facilitated instruction of AP Calculus AB and BC courses for 50+ high school students.
- Scheduled targeted tutoring sessions and office hours to improve comprehension of course material.
- Achieved average student scores of 4.81 in AP Calculus AB and 4.77 in AP Calculus BC.
Skills
- Machine Learning: Supervised and Unsupervised Learning, Time-Series Analysis, Natural Language Processing (NLP), Deep Learning, Regression, Classification, Clustering, Neural Networks, Gradient Boosting (XGBoost, LightGBM), ARIMA
- Data Analysis & Modeling: Statistical Modeling, Experimental Design (A/B Testing), Quantitative Research, Data Visualization (Matplotlib, Seaborn, ggplot2), Exploratory Data Analysis (EDA)
- Programming Languages: Python (Pandas, NumPy, Scikit-learn, TensorFlow, PyTorch), R
- Research & Documentation: Research Reporting, Technical Writing, Data-driven Storytelling, Research Ethics, Documentation, Manuscript Preparation
Accomplishments
Developed and trained a time series model to forecast 1-day and 3-day Bitcoin prices
- Compared the performance of two distinct modeling approaches: the classical ARIMA model and the machine learning-based XGBoost Regressor.
Developed an AI-powered email agent to automate the process of setting up networking coffee chats and cold emails with suitable professors and Ph.D. students.
- Utilized the LangChain framework to orchestrate a series of tasks, including scraping academic websites for contact information, analyzing researcher profiles to determine suitability, and dynamically generating personalized email templates.
- Implemented a Large Language Model (LLM) to understand the user's intent and generate compelling email content, ensuring each message was tailored and contextually relevant.
Affiliations
- Performing Arts: My passions include music and dance. I am a classically trained pianist (ARSM Diploma in Music Performance with Distinction), a soprano in Berkeley Chinese A Capella, and a ballet dancer (RAD Grade 5 exam with Merit).
Timeline
Research Assistant
Berkeley AI Research
05.2025 - Current
Teaching Intern
BASIS International
08.2023 - 06.2024
Research Assistant
University of Science and Technology of China
07.2023 - 08.2025