Projects

Personal WebsitePermalink

• Developed a customized open-source static website powered by Jekyll and Minimal Mistakes
• Automatically enable dark/light themes based on the user’s OS/browser preferences with the prefers-color-scheme CSS media feature
• Created custom containers for a grid view publications page using Liquid, HTML, and CSS

Distributional Deep Q-learningPermalink

• Implemented Google’s C51 algorithm in PyTorch, and trained it on various Atari games such as Pong and Breakout
• Exploited high performance GPU computing to run highly parallelized, batch jobs on the Blue Waters supercomputer
• Programmed single-layer networks from scratch and tuned hyperparameters for CNN, RL, and NLP deep neural networks

Modeling and Evolutionary OptimizationPermalink

• Simulated slithering snake and elastocapillary fins using 1D Cosserat rods and position-Verlet time integration in Python
• Modeled muscular activations, elastocapillary forces, friction and self-contact boundary conditions
• Implemented covariance matrix adaptation evolution strategy (CMA-ES) in Python to minimize generic loss functions
• Employed CMA-ES for 40 generations to maximize the snake’s forward velocity and achieve desired fin mechanics

Turbulence Compensation Table - University of Arizona Senior DesignPermalink

• Devised a two axis, active suspension table system for use in a super first-class airplane suite with five colleagues from various engineering disciplines. Awarded the Excellence in Aerospace Electronic System Design prize
• Utilized GD&T and manually machined over twenty moving parts for final table assembly
• Conducted rigorous trade studies and communicated limitations with sponsor to adjust design requirements