Mrunal J. Sarvaiya

I have built race cars, developed algorithms for mobile manipulators that clean hotel bathrooms at Peanut Robotics, and recently joined a PhD program at NYU to explore the integration of Model Predictive Control and Reinforcement Learning for aerial transportation systems.

People often ask me, “What made you pursue a PhD after developing planning and controls software for a robotics company for 4 years?” My intellectual curiosity bucket simply wasn’t satisfied. I found myself craving the fulfillment I felt when I dove deep into a specific topic simply because it intrigued me. I wanted to immerse myself in a specific niche and spend the next few years filling up my curiosity bucket. That’s what led me to pursue a PhD.

When faced with a decision with equally appealing options, such as starting a PhD program or continuing in the industry, I ask myself, “In 10 years, which choice will I regret not taking?” I knew for a fact that I’d regret not pursuing a doctorate.

What Do I Bring to the Table?

\(S = f_{\theta}(E, R, c)\)

I have developed a context-dependent (\(c\)) skillset (\(S\)) that sets me apart from the average roboticist and researcher. My research capabilities (\(R\)) facilitate my ability me to dive deep into algorithm details, critically question the validity of proposed methods and persistently solve hard problems. Meanwhile, my engineering expertise (\(E\)) allows me to intuitively understand real world requirements and limitations, forsee failures and systematically account for these considerations during system development.

In the controlled environment of an R&D lab or startup, a robot success rate of 90% may suffice. To enable robots to operator in the messy, unpredictable real world, we need 99.99%. I’ve optimized \(\theta\) so that I can seamlessly transition to play the role of a research scientist, a robotics engineer and everything in between. I aspire to lead teams to close the remaining 9.99% gap and drive integration of reliable robots into our everyday lives.