• Designed and built Open When, a letter-writing web app using React frontend and Supabase backend

• Developed Gratie, a stack-based esoteric programming language interpreter using Rust

• Courses: Deep Reinforcement Learning (in progress), The Missing Semester of Your CS Education (completed)

• Developed a Julia-Rust software tool that enables seamless model-in-the-loop and software-in-the-loop simulation between the P2 drone’s Julia plant model and its onboard Rust controller; eliminated over 80% of code iteration overhead and saves the team >15 hours/week

• Built model validation tool in Julia to quantify and visualize model mismatch in drone actuation forces and passive aerodynamic forces from flight test data; tool is being integrated into team's model development workflow as a key validation step

• Developing deep learning approach to predict solution cases for optimization-based trajectory generation in C++

• Implemented deep imitation learning neural network in Python using PyTorch for efficient motion primitive sampling in time-optimal long-range planner using ROS; created dataset, designed architecture, tuned hyperparameters

• Built greedy algorithm that performed 200x faster with only a 5% longer trajectory time for high-acceleration vehicles

• Implemented jump-point search algorithm and ran ablation studies in order to compare performance to A* in planner