About
LinkedIn
GitHubHi! I am Mingjian Li (李明键).
I am currently a senior student at NYU Tandon, majoring in Computer Science and minoring in Mathematics. I am passionate about and have experience in 3D Reconstruction, Full-stack Development, Machine Learning, and Data Science.
Before I transferred to NYU, I was studying Accounting at SUFE.
I love cats, hiking, photography, board games, and ping-pong!
Education
New York University
Jan. 2024 - May 2026 (Expected)
B.S. in Computer Science ; Minor in Mathematics
GPA: 3.983 / 4
- Worked as a peer tutor for Calculus III at TRIO program.
Courses
| Course Name | Grade |
|---|---|
| Machine Learning | 4.0 |
| Operating System | 4.0 |
| Design and Analysis of Algorithms | 4.0 |
| Object Oriented Programming | 4.0 |
Shanghai University of Finance and Economics
Sep. 2022 - Dec. 2023 (transferred out)
B.S. in Accounting ; Minor in Statistics
GPA: 3.983 / 4
- Renmin Scholarship First Prize
- Second Prize in China Undergraduate Mathematical Contest in Modeling Shanghai District
Research
Research Areas
3D Reconstruction
Computer vision techniques for 3D modeling and reconstruction from images, focusing on autonomous navigation and augmented reality applications
MRI & Deep Learning
Medical imaging research focusing on fastMRI techniques for accelerated magnetic resonance imaging, reducing scan times while maintaining image quality
Research Experience
Dynamic 3D Scene Reconstruction with 3D Gaussian Splatting
Research on building a pipeline for reconstructing dynamic 3D scenes from multiple camera viewpoints.
- Ongoing output: a benchmark dataset with multiview, multiframe human dancer sequences, including building a full pipeline for raw data processing, 3D Gaussian construction, mesh reconstruction, and parameter tuning.
- Replicating and comparing multiple 3D Gaussian methods (Dynamic Gaussian, 4DGS, Gaussian Surfel, 2DGS, SUGAR, etc.) on the dataset, achieving high-quality visual outputs (PSNR > 35).
- Evaluating mesh reconstruction techniques such as SDF, Poisson reconstruction, Marching Cubes, and tetrahedralization to improve geometric consistency.
- Benchmarking human body models including SMPL, supported by OpenPose keypoint detection, to assess accuracy and robustness for pose estimation tasks.
Research Skills
Programming
Python, C++, MATLAB
Frameworks
PyTorch, OpenCV, NumPy
Medical Imaging
fastMRI, k-space
Deep Learning Models
Diffusion, Transformer
Internship
Software Engineer Intern
SeeM(useums) – Pittsburgh
Dec 2024 – May 2025
- Developed a front-end application using React, Next.js, and Redux, featuring authentication, project management, and integration with an AI agent, deployed on AWS for scalable infrastructure.
- Contributed to back-end development by designing data schemas and setting up environments using Docker with Flask in Python, ensuring efficient, scalable, and consistent deployment.
- Implemented a 3D reconstruction pipeline using 3D Gaussian splatting and SuGaR technology, featuring CUDA extensions and support for multiple data formats. This setup consistently achieves a steady quality with PSNR over 26, and includes front-end visualization using Three.js.
Quant Research Consultant
WorldQuant Brain – Remote
Sep 2023 – Jan 2024
- Build market-neutral, medium-frequency strategy based on cross-field data and sentiment analysis.
- Implement backtesting and refine strategy to improve Sharpe ratio, turnover, and fitness performance.
Audit Intern
KPMG – Shanghai, China
Jan 2023 – Feb 2023
- Recognized risk points and conducted assessment in collaboration with colleagues to enhance analysis procedure.
- Performed inquiry, recalculation, observation and other substantive procedures for the integrity of assets, payroll and expense accounts and contributed to regulatory compliance audits for 5 international companies.
- Investigated subsequent events and conducted substantive tests of year-end transaction for each corporation.
Projects
Featured Projects
Kaggle: Jane Street Real-Time Market Data Forecasting
- Conduct data-cleaning, pre-processing, unsupervised learning including clustering and PCA, and visualizations
- Implement multiple feature transformations and applied regularizations for better performance。
- Fit different models, including SVM, CNNs, RNNs, LSTM and ARIMA with multiple sampling methods
Portfolio Management and Analysis App with AI
Modern, responsive portfolio website built with Next.js and Tailwind CSS. Features a clean design with smooth animations, responsive layout, and professional presentation of skills and projects.
- Tools Used:Bert, Transformer, OpenAI API, React, ExpressJS, MongoDB, NodeJS, Redux
- Developed a portfolio management app with a visualization dashboard for tracking holdings and market trends, integrated Large Language Models and News API for sentiment analysis, delivering tailored financial advice.
Full-stack Web App with AI Chat Box: EcoPantry
- Tools Used: JavaScript, React, ExpressJS, MongoDB, NodeJS, Azure API, Redux, Sass
- Developed a food management app with user-authentication, multiple pages and sidebar navigator featured with an AI chat box to analyze food and user health data and give management, health and recipe suggestions
- Implement a Dashboard to summarize food storage information with a searching tool and a updated list.
- Collect data and store hashed user information and formatted food information on a MongoDB database.
Technical Skills
Programming
Python, C++, JavaScript, TypeScript
Web Development
React, Next.js, Node.js, HTML/CSS
Machine Learning
PyTorch, TensorFlow, Scikit-learn
Tools & Others
Git, Docker, AWS, Linux