About Me
I am a Master of Science student in Data Science at Harvard University (September 2025 - December 2026) and a recent graduate with a Bachelor of Arts in Mathematics and Computer Science from New York University (September 2022 - May 2025). My research interests focus on 3D reconstruction, rendering-induced intelligence, and generative 3D models, particularly in applications to computer vision and robotic systems. Previously, I worked as a Deep Learning Researcher at NYU Tandon AI4CE Lab and as a Data Science Researcher at NYU Langone Health Radiology Research Department, where I contributed to projects in 3D reconstruction, assembly, and medical imaging analysis. Currently, I work as a Deep Learning Researcher at Harvard & MIT AI and Robotics Lab.
Available: May 26th, 2026 - August 30th, 2026
Target Roles: ML Engineer, Computer Vision Engineer, Data Scientist, Software Development Engineer, Quant Researcher
My Edge: My interdisciplinary background in Mathematics, Computer Science, and Data Science provides a strong foundation for technical problem-solving across domains. I bring extensive experience in deep learning frameworks (PyTorch, TensorFlow), cloud infrastructure (GCP, AWS), and production ML systems, along with proven research capabilities demonstrated through publications at leading conferences. Whether optimizing model performance, building scalable data pipelines, or developing financial models, I thrive in translating complex technical concepts into impactful solutions.
🔥 News
- 2025.06: 🎉🏝️ GARF accepted to ICCV 2025! See you in Hawaii! 🌺
- 2025.02: 🎉🎉 Accepted to Harvard University's Master of Science program in Data Science!
📝 Publications
PAGE-4D: Disentangled Pose and Geometry Estimation for 4D Perception
Kaichen Zhou, Yuhan Wang, Grace Chen (2nd author), Xinhai Chang, Gaspard Beaudouin, Fangneng Zhan, Paul Pu Liang, Mengyu Wang. Project / Paper
GARF: Learning Generalizable 3D Reassembly for Real-World Fractures
Sihang Li, Zeyu Jiang, Grace Chen (2nd author), Chenyang Xu, Siqi Tan, Xue Wang, Irving Fang, Kristof Zyskowski, Shannon P. McPherron, Radu Iovita, Chen Feng, Jing Zhang. Project / Paper
💻 Research Experience
Harvard & MIT, AI and Robotics Lab
June 2025 - Present, Deep Learning Researcher
- Co-developed selective layer fine-tuning approach for VGGT transformer optimizing hyperparameters
- Implemented evaluation metrics including DTU accuracy, ATE/RPE for pose, and Occlusion Accuracy tracking
- Enhanced SFTTrainer developing custom classification loss functions improving prediction accuracy by 75%
- Evaluated performance using confusion matrix analysis improving class-wise precision with F1-score metrics
NYU Tandon, AI4CE Lab
August 2024 - Present, Deep Learning Researcher
- Processed 7 archaeopteryx thoracic and spinal bone fragments achieving assembly using PuzzleFusion and models
- Fine-tuned model parameters on archaeological datasets using PyTorch achieving 25% improvement in accuracy
- Developed BaggingGrid simulation pipeline using PyTorch and trimesh for training data generation processes
- Created 50% more varied training datasets for model robustness handling mesh processing and fragmentation
NYU Langone Health, Radiology Research Department
February 2024 - May 2025, Data Science Researcher
- Built analysis framework for 1.2K+ patient datasets supporting clinical research workflows with pipelines
- Developed data cleaning pipelines using Python, Pandas, and NumPy for radiology report processing
- Deployed Large Language Models for radiology report processing and cyst characteristic extraction from records
- Conducted statistical analysis on 200 radiology reports using Cohen's Kappa hypothesis test with R
💼 Professional Experience
Guotai Jun'an Securities
May 2023 - July 2023, Investment Banking Intern, Shanghai, China
- Built data engineering pipelines for 10+ client financial analysis projects with ETL processes automation
- Set up MySQL databases automatically pulling data from multi-source financial systems for deal evaluation
- Conducted data analysis and preprocessing using Python, Pandas, and NumPy for market research
- Implemented data cleaning procedures and missing value imputation processing financial statements for pitch books
🎓 Course Project
StyleMe: AI-Powered Fashion Styling Assistant
Technologies: Python, Docker, MLflow, FashionCLIP, GCP, MLOps
- Built MLOps infrastructure using Docker containers streamlining deployment model reproducibility for production
- Developed GPU-accelerated training pipelines for FashionCLIP models on NVIDIA V100 GPU in GCP
- Integrated MLflow tracking improving model convergence speed by 60% through hyperparameter tuning
- Implemented automated model versioning monitoring achieving 98% uptime for fashion recommendation service
🛠 Skills
Programming Languages
Python, SQL, Java, R, Spark, C, Bash, Shell, HTML, CSS, JavaScript
Machine Learning
PyTorch, TensorFlow, Scikit-learn, Hugging Face, Transformers, OpenCV, XGBoost, MLOps
Data Analysis
Pandas, NumPy, Scipy, Matplotlib, Seaborn, A/B Testing, Statistical Analysis, Hypothesis Testing
Tools & Technologies
Docker, Git, Jupyter, Tableau, Looker, Airflow, DBT, ETL Pipelines, GCP, AWS
Languages
Mandarin Chinese (Native), English (Fluent)