Pranav Rajaram
Data Science, UC San Diego
"Creative data scientist with strong fundamentals in machine learning, software engineering, and AI development. Shows high upside in predictive modeling and full-stack development environments, with a knack for data visualization and clear communication."
Game Film
Highlight reel of key projects showcasing technical skills and problem-solving abilities.
Developed a machine learning pipeline on MiLB tracking data to model baserunning with logistic regression and optimize cutoff play decisions with a Random Forest classifier, visualized through an interactive dashboard.
Built an interactive NFL guessing game inspired by Wordle, challenging players to identify NFL players based on attributes like team, position, and stats. Features dynamic feedback and player database integration.
Utilized XGBoost to model NFL tracking data, assigning tackle probabilities to plays to develop a new 'Tackle Rate over Expectation' metric. Commended for quality of visualizations and animations by NFL analytics departments.
Used Aerial and Terrestrial Laser Scanning data to build 3-D Canopy Height Models, generate engineered treelists, and train machine learning models to predict tree species and forest attributes for wildfire simulation.
Built a Random Forest Regression model with scikit-learn to analyze a dataset of power outages, used techniques like one hot encoding, hyperparameter tuning, and cross-validation to optimize model predictions.