Varun Ganjigunte Prakash

Bengaluru · India
Hi! I'm Varun.
I am a Computer Vision engineer and Robotics practitioner. I am gulping the concepts to learn, to fail, and then to make ideas work by experience since I believe that practice brings real knowledge. Robots fascinate me and I enjoy building them. Building goal-oriented machines that can sense, plan, act safely, and help people is my long-term life goal. I also like to work on agricultural robotic systems for helping our farmers, making intelligent machines as accessible and affordable as cell phones. I am self-driven and chasing my vision. I pitch in to help projects and people, if it works I am lucky to see their happy faces.


I completed my undergrad at Sri Jayachamarajendra College of Engineering at Mysuru. I am passionate about solving problems in the intersection of Computer Vision and Robotics. I have spent time working on computer vision and robotics challenges.

My brief bio:
Machine Learning Engineer - Computer Vision
03.2021-current      CogniAble      Bengaluru, India
Engineer
06.2018-03.2021      L&T Technology Services Limited      Bengaluru, India
Computer Vision Intern
07.2018-10.2018Skylark DronesBengaluru, India
Embedded Systems Intern
06.2016-12.2016Logichive Solutions Private LimitedMysuru, India




Projects


6–DOF pose estimation, planning and control


Aim is to design, simulate in ROS and testing on custom built manipulator for 3D object grasp with Deep Learning
Platform : ROS, Gazebo, TensorFlow




Indian currency recognition application and food classifier application


1) Demonstration of Cash Recognition mobile application for Indian currencies
2) A food classifier mobile application which identifies and discovers related information about the food such as nearest restaurants where it is available, the item cost etc., on the click of a photo. Developed for 20 most common Indian food items
Platform : TensorFlow (Keras), Python, Android




Dexterous Service Robot


The purpose of our project is to build a home assistant robot to assist differently-able and aged persons. The proposed robot helps such people by performing some of the common tasks involved in our daily life through human-machine cognitive learning. A 5 DOF Dexter ER-2 arm mounted on a vehicle is operated by voice commands to selectively search and deliver the article required by the user. Simulation is also performed to check the feasibility of the planned path for arm's joints and to avoid a collision in a dynamic 3D environment
Platform : ROS, Python




Colour based object sorting using DEXTER ER-2


The aim of this project was to sort objects based on color by using a heavy duty robotic arm (DEXTER ER-2). The algorithm used Image Processing and Inverse Kinematics concepts. Colored objects of different sizes were also sorted
Platform : MATLAB




Autonomous object delivery robot


The project aims at selective object delivery based on shape, size, and color of the objects. Objects and surrounding obstacles were analyzed by using Image Processing concepts programmed in OpenCV-Python. Fire Bird V robot was navigated using XBee wireless communication. A robotic gripper has been designed to pick and deliver objects. The project helps to understand and improve the autonomous delivery robot system which is efficient and self-reliant
Platform : AVR Studio and Python


Interests

Computer Vision
Robotic Manipulation
Autonomous Mobile Robots
Deep Learning
Multi-Task Deep Reinforcement Learning
Robotic System Design
Control Theory
Machine Learning for Computer Vision



Publications:
1) Autonomous Service Robot

AIR 2019: Proceedings of the Advances in Robotics 2019


Curriculum Vitae

Please refer to my CV for more details, Download CV