UPenn CS Master AI PM
I design AI-powered products using LLM workflows,machine learning systems, and full-stack engineering.
Projects Gallery
YiYi AI Closet
YiYi AI Closet is an AI-powered virtual wardrobe that helps users organize their clothing and generate personalized outfit recommendations with visual try-on.
Music Style Converter
Machine learning project that converts traditional Chinese music into jazz, using generative models and spectrogram feature engineering.
Working Experience
Aethorion AI
Sep. 2025 — Nov. 2025AI Researcher
- Managed the integration and training of multimodal models on AWS.
- Benchmarked Empirical models (XGBoost) against Multimodal models.
- Confirmed a 20-30% error reduction in cement strength forecasting.
Capgemini
Jun. 2024 — Sep. 2024Data Scientist Intern
- Built LLM-based framework for automated review classification.
- Used prompt engineering and feasibility testing for text extraction and classification.
- Improved classification recall from 60% to 90%.
ISF Research Institute
Jan. 2024 — Mar. 2024Quantitative Associate Analyst
- Supported multi-factor quantitative model with data cleaning (Pandas).
- Built IC trend visualization with Matplotlib for factor effectiveness assessment.
- Delivered analysis and visualization outputs to support research decisions.
Tongsheng Asset Management
Jul. 2023 — Aug. 2023Project Assistant
- Conducted market research and reporting.
- Comparative analysis of GPU products and IDC options via Excel.
- Supported investment decisions with data-driven insights.
Nielsen
Dec. 2022 — Jan. 2023Data Analyst (PTA)
- Assisted mentor to conduct strategy analysis for 2 major clients.
- Integrated 500+ questionnaires and competition data.
- Built Excel visualizations for brand awareness and conversion.
Education
Master of Computer and Information Technology
University of Pennsylvania, Philadelphia PA
Aug. 2025 - May. 2027
B.A. in Business Economics & B.S. in Mathematics
University of California, Irvine CA
Sep. 2021 - Jun. 2025
Awards & Hackathons
AETHORIX V1.0: An Integrated Scientific AI Agent for Scalable Inorganic Materials Innovation and Industrial Implementation
Aethorix v1.0, an integrated AI agent that combines scientific corpus reasoning, diffusion-based inverse design, and machine-learning interatomic potentials to optimize inorganic materials such as cement in real industrial workflows.
Personal Role — AI Researcher
Focused on Validation & Implementation
- Pipeline Execution: Managed the integration and training of multimodal models on AWS using combined industrial and simulation datasets.
- A/B Testing: Benchmarked Empirical models (pure XGBoost) against Multimodal models to quantify the impact of physics-informed features.
- Result Validation: Confirmed a 20-30% error reduction in cement strength forecasting, validating the effectiveness of the integrated AI workflow.