Stanford NeuroAILab2018 - Present
Used model-free deep reinforcement learning (A3C, PPO, DQN) to solve visually rich 2D tasks. Conducted an architecture search over pre-trained convolutional network features for fine-tuning. Adapted deep reinforcement learning methods for 3D continuous control and navigation tasks with visual input and a continuous action space.
UT Neuroimaging and Electrophysiology LabJune 2017 - present
Used machine learning, digital signal processing, and time series analysis techniques on ECoG data to find significant patterns in the usage of critical language production areas of the brain. Methods include: Support Vector Machines, Spectral Coherence, Granger Causality, Wavelets, Fourier Analysis, Filtering, Clustering, Phase Locking Value, Phase Slope Index, etc
Rice Digital GymFall 2016 - present
Developed an ionic mobile application and a node.js backend that is currently deployed in AWS
North Carolina State University Power AmericaJune 2015 - December 2015
Data mining and web scraping on the wide-bandgap semiconductor product space. Created a tool for visualizing the WBG semiconductor market and associations with academic researchers.