Welcome! I am Siddharth Saha.
I graduated from the mecca of robotics, The Robotics Institute @ CMU where I worked on quadruped robot. I was in the M.S. Robotic Systems Development program, Batch of 2024. My program’s cornerstone was the 1.5-year-long capstone project, which I was fortunate to have with Prof. Sebastian Scherer, AirLab. I used the Unitree quadruped robot for search and rescue in unknown and cluttered environments. Through my life’s work, I want to advance robotic capabilities in unstructured environments with three focuses: motion planning, legged locomotion, and optimization.
I graduated from IIT Bombay in 2021, attaining a Bachelor’s (with Honors) in Mechanical Engineering and a Minor degree in the Department of Computer Science. During my undergraduate, I was fortunate to work under the guidance of wonderful mentors. I worked with Prof. Leena Vachhani on my Undergraduate Thesis and Prof. Prasanna Gandhi on another research project related to modeling ultra-flexible inverted pendulum
At the institute, I was the founding member of a fully-funded technical team building a quadruped robot and motivated to solve the international urban search and rescue challenge RoboCup Rescue League
My interests lie at the intersection of optimization and planning for legged robots.
What distinguishes me, however, is my resilience, curiosity, and hacker mindset.
Resilience: Despite my humble family background, I was dead set on a career in robotics during college. I worked for a year after graduation in finance to fund my M.S. tuition. I simultaneously pursued robotics research and completed my Google Summer of Code at JdeRobot.
Curiosity: I take delight in questioning what goes behind everyday technologies that we take for granted.
Hacker mindset: My dopamine hits arise from hacking around problems. For instance, I use TeXstudio over OverLeaf because of its speed. But TeXstudio’s symbols packages clashed with the powerline symbols on my Linux rice. I put my TeXstudio inside a Docker environment and faced no problem ever again! I have blogged about the steps to achieve this.
My software engineering experiences range from short-life codes, which won me national and international robotics competitions, to long-life codes, contributing to the mammoth code base at Goldman Sachs and doing open source at JdeRobot.
Besides my love for coding and robotics, I enjoy admiring talent across arts and music — to be inspired by people more gifted than me. Watching Matt McGuire at the drums never fails to electrify me, and listening to Coldplay fills me with joy.
This website is a quick outlet for my thoughts and life updates
You can also read more about my professional experience & projects
Questions That Inspire Me
- Multi-modality during Legged locomotion:
- What gait patterns could be more energy efficient than their wheeled counterparts?
- How does the control generalize over complex foothold constraints, like steep stairs, slippery terrain, sharp rocks, rubble?
- Can we design perception/sensing algorithms to infer foothold constraints from the world in real-time?
- When a biped has lost balance and is about to fall, can it reflexively extend its arms to brace for impact or reflexively grip a rigid object to prevent the fall?
- Optimization
- How can we combine methods from Trajectory and Gait optimization, Pose Graph optimization, and Policy optimization to answer the above questions?