Machine Learning Intern
Gilroy, California, United States
Worked alongside Prof. Paaras Chand to explore Deep Reinforcement Learning.Self-studied Berkeley grad courses, covering the mathematical theory, implementation, and real-world issues of 10-15 deep RL techniques. Included:Deep NNs and their use in Policy Gradients, A2C algorithms, Deep-Q Learning, TPRO and PPO, Value Networks, On/Off -policy/Offline Learning, Exploration, Multi-agent / Self-play training, Markov Decision Processes, Monte-Carlo methodsDeveloped and trained 6-8 PyTorch models and compared their performance in zero-sum games (Connect Four); utilizing Convolutional Neural Networks, multi-agent / self-play trainingUsed ML and application deployment technology, like PyTorch, Gymnasium, OpenSpiel, NumPy, Jupyter Lab/Colab, Linux and CUDA, SSH, pygame, ExpressHelped other groups within Gavilan working on physics research to incorporate neural networks (PINNs) into their researchGathered findings into a poster and video & presented at a large STEM symposium