Rohan M

Computer Science, Machine Learning,
Computer Vision, Music, Philosophy,
Applied Math



Someone once said that your résumé is just a documentation of all your sufferings till now. Here is my list.

  • Graduated from college in Mumbai, India. Learned enough to know that I knew nothing about computer science or math, but enough web development to be able to build this website.

  • Joined AitoeLabs, a computer vision startup, without knowing what computer vision was. Trained some interesting models and built some cool products, became the Engineering Lead of the company and helped them win $1.4M ARR. All of it while working with super talented friends.


  • Found out that I liked running large scale experiments, so I joined the Computer Vision research group at IIT, Bombay with my AitoeLabs friends. Starting with zero experience in writing papers or running experiments (and LaTeX), spent countless sleepless nights over a period of 2 years with Suraj Kothawade learning every possible Shell scripting hacks, enough GPU skills and machine learning/computer vision hyperparameter hacks to publish some of our works at WACV '19.

  • As a chance to get over a case of impostor syndrome, I started giving a series of talks on how to practically get started with computer vision research. Mentored quite a few students who attended the talks with their research projects later.

  • Moved to Munich, Germany. Struggled to buy groceries but worked on making reading books more interesting with great people at the Digital Product School, UnternehmerTUM and built a cool app (Bukdu, available on the Android Play Store).

  • Started grad school at the Courant Institute of Mathematical Sciences, New York University in 2019. Did some research stuff. Attended lectures and seminars by researchers who got me interested in computer vision.

  • Joined the computer vision team at Pinterest where I work on addressing the problem of inclusion and diversity in visual classifiers and help reduce socio-cultural biases from the models deployed on a billion-image scale.
  • I also work on making content more conducive to inspiration through generative models and understanding and manipulating entities in images. My goal with this work is to let people visualize a portion of what they imagine/dream when taking on creative projects by letting them alter/tamper objects as they like.