I have a background in data science, mathematics, experimental physics, robotics, and applied engineering, with hands-on experience in Python, SQL, C++, machine learning, quantitative modeling, data visualization, ROS 2, and simulation.

My GitHub focuses on practical data, modeling, and engineering projects, including used-car price modeling, autonomous racing simulation, experimental physics data analysis, computer vision, API-based analytics, and biostatistical modeling.

A machine learning project for analyzing used-car depreciation patterns and estimating resale prices from listing-level vehicle attributes.

This project builds a practical valuation workflow, including data cleaning, exploratory analysis, feature engineering, LightGBM modeling, benchmark comparison, segment-level error analysis, and a Streamlit app for interactive vehicle price estimation.

Highlights:

• Cleaned and processed large-scale used-car listing data
• Engineered depreciation-related features such as vehicle age, mileage transforms, and engine displacement
• Trained a LightGBM regression model for resale price prediction
• Benchmarked model performance against simpler baseline models
• Built a Streamlit app to demonstrate price estimation and depreciation curves

Tech Stack: Python, Pandas, LightGBM, Scikit-learn, Streamlit, Matplotlib, Regression Modeling

A ROS 2 autonomous racing simulation project using LiDAR-based local planning in the F1TENTH simulator.

This project combines wall following, follow-the-gap, local trajectory planning, predictive speed control, racing-line biasing, and ML-assisted tuning. It demonstrates how robotics simulation and data-driven parameter tuning can be used to improve autonomous vehicle behavior.

Highlights:

• Built a LiDAR-based local racing planner for the F1TENTH simulator
• Combined wall following, follow-the-gap, and local planning logic
• Added predictive speed control based on curvature, clearance, braking distance, and steering smoothness
• Used parameter tuning and ML-assisted adjustment to improve simulated driving performance
• Organized configs, scripts, models, and test results for reproducible evaluation

Tech Stack: ROS 2, Python, F1TENTH, LiDAR, NumPy, YAML, Simulation

A portfolio of selected experimental physics work from the University of Michigan PHYSICS 391 laboratory course, focusing on quantitative data analysis, parameter estimation, curve fitting, and physical interpretation.

The work includes experiments involving the Hall effect, semiconductor conductivity, high-temperature superconductivity, spectroscopy, radiation, and other modern physics phenomena.

Highlights:

• Analyzed Hall voltage under varying current and magnetic-field conditions to investigate carrier behavior
• Analyzed temperature-dependent semiconductor conductivity and used quantitative fitting to estimate the germanium band gap
• Evaluated high-temperature superconducting behavior using resistance-temperature measurements, Meissner effect observations, and transition-temperature analysis
• Applied linear and nonlinear fitting, parameter estimation, statistical analysis, and scientific visualization to experimental data
• Documented experimental methods, analysis, and findings through technical laboratory reports

Tech Stack: Excel, Excel Solver, Nonlinear Least-Squares Fitting, Parameter Estimation, Experimental Data Analysis, Scientific Visualization

A computer vision project for traffic sign detection using YOLOv8 and custom lightweight architecture variants.

This project compares a standard YOLOv8 baseline with attention and feature-fusion extensions such as CBAM, LCFE, IMCMD, and YOLO-TS-style fusion. The goal is to improve small-object traffic sign detection and evaluate model robustness across day and night driving conditions.

Highlights:

• Trained and evaluated YOLOv8-based traffic sign detection models
• Compared baseline, attention-based, and feature-fusion model variants
• Evaluated mAP, parameter efficiency, and day/night robustness
• Organized experiment configs, scripts, reports, and result summaries for reproducible analysis

Tech Stack: Python, PyTorch, Ultralytics YOLOv8, OpenCV, Object Detection, Computer Vision

A Python data analysis project comparing 2020 and 2021 year-end music trends using Spotify API data, SQLite, SQL queries, and Matplotlib visualizations.

Highlights:

• Collected playlist data from the Spotify Web API
• Stored artist and track records in SQLite
• Used SQL joins to compare artists across years
• Identified artists appearing in both years and artists unique to each year
• Generated visualizations for artist overlap and trend comparison

Tech Stack: Python, Spotify API, SQLite, SQL, Matplotlib, JSON

A Python-based mathematical biology project that simulates HIV infection dynamics and immune response using ordinary differential equations.

Highlights:

• Modeled interactions between healthy CD4+ T-cells, infected cells, virions, and immune response cells
• Used SciPy ODE solvers for numerical simulation
• Compared immune response behavior under different parameter settings
• Visualized CTLp and CTLe trends over time
• Compiled results into a research report and presentation

Tech Stack: Python, NumPy, SciPy, Matplotlib, ODE Modeling, Biostatistics

A semester-long C++ programming portfolio covering console applications, games, ciphers, object-oriented design, and simulation.

Highlights:

• Built projects including a focaccia calculator, Rock-Paper-Scissors game, cipher program, Battleship game, and elevator simulation
• Practiced C++ fundamentals, functions, loops, strings, file input, and class design
• Developed larger multi-file programs using object-oriented programming

Tech Stack: C++, Object-Oriented Programming, File I/O, Console Applications

Programming: Python, SQL, C++
Data Science & Analytics: Pandas, Scikit-learn, LightGBM, SQLite, Matplotlib, Streamlit, API Data Collection, Data Cleaning
Machine Learning: Regression Modeling, Model Evaluation, Feature Engineering, Benchmarking, Error Analysis
Scientific & Experimental Analysis: Experimental Data Analysis, Parameter Estimation, Linear & Nonlinear Least-Squares Fitting, Numerical Analysis, Excel Solver
Computer Vision: PyTorch, Ultralytics YOLOv8, OpenCV, Object Detection
Robotics & Simulation: ROS 2, F1TENTH, LiDAR, Odometry, Ackermann Drive, Simulation
Tools: Git, GitHub, Ubuntu, WSL2, VS Code

I am currently building projects at the intersection of data science, business analytics, machine learning, scientific computing, and engineering applications. My strongest interests are data analytics, applied modeling, business intelligence, data-driven engineering, and practical quantitative methods for solving real-world problems.