Under the mentorship of Prof. Koushil Sreenath, I am developing of a machine learning policy using Proximal Policy Optimization (PPO). This project is conducted within the advanced NVIDIA Omniverse Issac Gym Environment, where I am focusing on training the Agility Robotics’ Digit, a 36 Degree of Freedom (DoF) humanoid robot, to function as an effective boxer and boxing trainer.• Completed starter projects such as training the Unitree A1 quadruped and the Agility Robotics’ Cassie bipedal robot to achieve stable walking. Designed and tuned various reward functions and implemented a random push function successfully.• Enhanced the robustness of the Cassie walking gait for real world applications by incorporating uneven terrain and adding observation and process noise. This is also being applied to Digit.• Currently working on training digit to talk and on improving our physics model of digit to the Sim2Real transition. The next phase of the project will focus on integrating new reward mechanisms aimed at stimulating movements that emulate boxing techniques such as punching and dodging.

- Founded Jews in Science and Engineering, at UC Berkeley, a club that empowers students to create engineering projects for the betterment of the Jewish community and the Bay Area.- Hosted a 3D modeling workshop to teach students the basics of Computer Aided Design, through the development of a fun and relatable project

Completed a 5 week course on Nasa subject matter. Researched and designed a rover to land on the moon. For the virtual onsite experience I worked with a team to design a mission to build a habitat on the moon.