Siddharth Girdhar

Bio

I am a senior studying computer science at Carnegie Mellon University and a software engineering intern at Meta.

My research interests are in using reinforcement learning to learn generalizable and robust policies for robotic manipulation.

In my free time I enjoy sports, travelling to new places and trying new cuisines

Download my resumé.

Interests

Education

B.S. in Computer Science (Concentration in Machine Learning), 2022

Carnegie Mellon University

Software Engineering Intern

Jun 2022 – Present Menlo Park, CA

Research Assistant

Jan 2021 – Present Pittsburgh, PA

Exploring the application of information compression in reinforcmenet learning algorithms for learning robust policies for robot manipulation tasks
Investigated the use of disentangled representation learning for dynamics model to see if it leads to better generalization
Developed a matrix‑based toy environment in OpenAI Gym to stress test causal reasoning algorithms

Software Engineering Intern

Sep 2020 – Jan 2021 Pune, India

Led a team of three to design and implemented a fault tolerant controller for a quadcopter in simulation that prevents crashing after single motor failure.
Implemented safety compliance software in the PX4 firmware to satisfy government regulations such as avoiding no‑fly zones and verifying digital certificates

Software Engineering Intern

May 2020 – Aug 2020 Pittsburgh, PA

Developed a tool in C++ that analyzes quantitative and visual analysis of the differences between inferred ground generated from point cloud data
Analyzed pipeline that generates inferred ground from point cloud data to identify and eliminate unneccesary steps, optimizing for runtime and quality of output.

Research Assistant

Oct 2018 – Oct 2019 Pittsburgh, PA

Built a restaurant simulator using V‑REP to use as a testing platform for robot waiters
Designed and ran a Human‑Robot Interaction study that used eye gaze information from participants to predict their intent and use robot gestures to manipulate it.