About Me

R S Nikhil Krishna .

Why do I do what I do? Because "You either die a programmer, or live long enough to be the programmed". I am a student at IIT Madras, programmer, numismatist, Music Lover (❤) and gamer. Current fields of interest include Computer Vision, Machine Learning, Reinforcement Learning, and AGI

My Career

Detect Technologies

In Detect Technologies, I lead the Computer Vision and Artificial Intelligence applied to Analysis of Drone Acquired Data.

Computer Vision Lead & Data Analyst

Computer Vision Group, CFI

Headed the Computer Vision Group at CFI, IIT Madras for a year, paving way for innovative projects in the field of Deep Learning and working closely with industries to solve challenges posed by them

May 2017
Club Head

Bhabha Atomic Research Centre

I worked on biomedical computer vision, image segmentation, and render optimization

Jun 2016
R&D Trainee

IIT Madras

I now study at IIT Madras, Chennai, India. I am majoring in Mechanical Engineering, with a minor in Robotics

Undergraduate Education

My Skills

My Projects

Model Generalisabilitiy - Thompson Sampling on Disentangled Latent Variables

A project aimed at trying to improve the generalisability of a DQN across games using Visual Concept Learning to learn disentangled representations of certain state-spaces, and using this to build a complete dynamics model of a game

Intelligent Ground Vehicle

The IGV of IIT Madras is a self-driving vehicle that navigates through unfamiliar tracks, avoiding dynamic obstacles. I worked on using Deep Learning for lane discovery in the same and identifying statistically likely thresholds for the same.


I undertook Org-E as a project to build a solution to the clutter we buildin our digital life. It organizes files into folders based on type, and optionally rearranges the existing folders and the relocation of files too

Interactive Distortion Correcter

Finding the right value of the squared and fourth order coefficient for Brownian radial distortion correction can be a pain in the ass. This project was aimed at easing that process and helping you easily find the coefficient for a given image