Projects

My work can be categorised in two domains.
Besides below list, I have applied Machine Learning to many real world problems.

tl;dr:

I am really good at Machine Learning, Computer Vision and Computer Graphics.
I possess excellent algirithmic knowledge as well as application layer understanding.


Graphics

2015
  • Designed and developed Commercial Game to help children remember Number Tables
  • GPU Profiler for AMD, computing statistics at runtime. It emmensly helps in understanding resource utilization
  • Created Augmented Reality Demo using Kinect For Windows V2 and OpenGL
2014
  • Android Platform targetted. Multiple fancy shaders on Android using OpenGL ES 2.0 and NDK
  • Got opportunity to work on Intel's GPU device driver for Windows XP. Gained intrinsic knowledge about GPU.
  • Wrote minimalistic math library for Graphics Programming. Intended audiance was folks interested in learning 'Behind the Scene' maths
  • Started developing Rendering Engine for Android OpenGL ES 2.0
  • Developed a GPU powered PhotoViewer App for Android platform
2013

Computer Vision

2015
  • Gained in-depth understanding of many Machine Learning algorithms by implementing those algorithms myself
  • Learning by doing, applied those algorithms to well established applications like OCR, Face Recognition, etc.
  • Realtime object tracking using Kalman Filter. Demo available here, hand is being treated as object
  • Developed algorithm for Gesture Recognition on Android. Used PCA as feature extractor and SVM for classification
  • Invented algorithm for Transmission Line Detection in Real Time. Publication work is under progress
  • Real Time Action Recognition using Kinect, have significant findings. Publication work under progress
  • Demonstrated a method "Traffic Vehicles Classification using Image Processing" for very dense traffic
  • Consulted start up in solving Lung Cancer Cell Classification problem, using Machine Learning and Image Processing
2014
  • Got opportunity to work with Kinect. Studied both Kinect for Windows V1 as well as Kinect for Windows V2, but later on switched to V2
  • Developed a Gesture Recognition Library at Persistent LABS. Not only created a dictionary of unambiguous gestures, architectured and engineered it
  • Developed apps for Smart Interaction like Powerpoint and Windows Media Player automation using Gesture Recognition Library
  • Predicting needs ahead of time, completed a famous Machine Learning course from Coursera - by Andrew Ng
2013