Ajay U. Mandlekar – Projects

Open-Source Software



I led the development of robomimic, a framework for robot learning from demonstrations.



I helped develop and currently maintain robosuite, a robot simulation benchmark powered by the MuJoCo physics engine.

Undergraduate Research and Work Experience

Apple Logo 

Apple iOS Location and Motion SWE Intern

Summer 2015

I spent a summer working at Apple. It was my first experience working with deep learning models like CNNs. I was very fortunate to have Jason Gauci as my mentor - I learned a great deal from him. I also had the opportunity to present my work to Craig Federighi, which was awesome.

CHIC logo 

Caltech High-Speed Integrated Circuits

Summer 2014

I spent a summer working in the Caltech High-Speed Integrated Circuits lab under Professor Ali Hajimiri. I implemented a feedback loop consisting of both hardware and software interfaces, allowing for a wireless power transfer prototype. You can find a video about some of the latest developments on the project below.


SLAC logo 

SLAC National Accelerator Laboratory

Summer 2013

I spent a summer working at the SLAC National Accelerator Laboratory under Professor Tor O. Raubenheimer and Dr. Juhao Wu. I investigated whether harmonic lasing could be used to make the shift towards higher energy radiation. This was my first research experience - I was very fortunate to have Professor Tor O. Raubenheimer and Dr. Juhao Wu as mentors who took time to teach me how to approach difficult scientific problems.

[report] [publication]

Undergraduate Projects

MP3 Player Figure 

Blackfin MP3 Player

Winter and Spring 2014

I built an MP3 player over 2 quarters in sophomore year for our embedded systems project class. By far the most difficult thing I have ever had to do. It even almost worked. You can find an extensive report (aka the longest thing I've ever written) below. Thanks Glen!


Netflix Figure 

Netflix Challenge

Spring 2015

A couple of my friends and I competed took a project class where we had to essentially participate in the Netflix Challenge. We leveraged a Netflix dataset of user-movie ratings to build a recommender system. This was a great experience in implementing models from papers and working with large datasets. We placed 3rd in our class.


Function Generator Figure 

Analog Function Generator

Winter 2016

I built an analog function generator from scratch for a project in senior year! It can generate a square wave, a triangle wave, and a sine wave with different voltage amplitudes and frequencies.


Smart Dog Bowl Figure 

Smart Dog Bowl

Spring 2015

I built a dog bowl that can open automatically when a special IR tag (worn by a dog) comes close enough.