1. Satellite Imagery Platform: Built new system for faster access of areas of interest for users in the UI using Cloud Optimized Geotiffs and storage of locations using a quadkey based approach. The system also integrated into providing high quality data for data science downstream processes. 2. Acceleration of processes: Accelerated this new platform on GPU’s, and demonstrated a much faster version of the platform using multi-threading. 3. Datacube: Wrote code for storing time series data and vegetation indexes from multiple satellite data sources and storing these multidimensional array representations in a database for easy querying. Demonstrated this system using Folium and Leaflet.js on interactive maps. 4. Impact: Work is currently be
Undergraduate Researcher
International Institute of Information Technology Hyderabad IIITH
Apr 2019 - Apr 2021(2 years 1 month)
1. Self Driving Car - Mahindra Rise Challenge Built circuits for the self driving car to switch between autonomous and free modes without the use of CAN bus. - Implemented Frenet Frame based obstacle avoidance to go along with Lidar based Occupancy and Costmaps for motion planning. - Implemented different sensor fusion algorithms for finding GPS waypoints and scene understanding using LiDAR and stereo based approaches. 2. Collision avoidance framework: Building non holonomic collision avoidance for real-world indoor scenes using novel methods (accepted to ECC’21). 3. Stereo vision: Implemented many classical and deep learning stereo vision algorithms for use in practical scenarios. Tech: MATLAB, Python, ROS.
Deep Learning Researcher
Indian Institute of Science IISc
Jun 2018 - Sep 2018(4 months)
1. Research in Computer Vision at the Video Analytics Laboratory (Computation and Data Science Center) under Professor Venkatesh Babu, working for PhD student K Ram Prabhakar. 2. Problems tackled include deghosting of images and image segmentation using siamese neural networks and Mask R- CNN's.
Intern
Polaris Wireless
May 2015 - Jan 2016(9 months)
1. Intern under Mr. Anand Srinivasan, Associate Vice President, Global Logic as a high school student. 2. Trained in GPS, GPRS and IoT systems apart from communication concepts like OSI and SS7 protocols.
Education
verified_user
0% verified
Bachelor of Technology - BTech, Electrical, Electronics and Communications Engineering
International Institute of Information Technology
Jan 2017 - Dec 2021(5 years)
Bachelor of Technology - BTech, Electrical, Electronics and Communications Engineering
International Institute of Information Technology Hyderabad IIITH
Jan 2017 - May 2021(4 years 5 months)
N
Science with Computer Science (Physics, Chemistry, Mathematics, Computer Science
National Public School Indiranagar
Jan 2015 - Dec 2017(3 years)
S
Sishu Griha High School
Sishu Griha High School
Jan 2005 - Dec 2015(11 years)
Projects (professional or personal)
verified_user
0% verified
Awards
verified_user
0% verified
S
Commendable performance in Mathematics
Sishu Griha High School
Award for best performance in Mathematics for freshman and sophomore years of High School.
S
Award for creativity and contribution towards Science
Sishu Griha High School
Facilitated for my contributions towards Robotics, Computer Science projects during High School years
D
Delhi Public School Hackathon
2nd prize received for developing an AR cross platform application with realtime feedback. Developed using NodeJS, AweJS, and Ionic Frameworks
S
Commendable performance in Computer Science
Sishu Griha High School
Awarded for highest and consistent marks during freshman and sophomore years of high school.
Publications
verified_user
0% verified
I
Non Holonomic Collision Avoidance under Non-Parametric Uncertainty: A Hilbert Space Approach
IEEE European Control Conference
Feb 2021
We consider the problem of an agent/robot with non-holonomic kinematics avoiding dynamic and static obstacles. Additionally there may be bounds/constraints on the configurational space of the robot in the form of lane/corridor boundaries. State and velocity noise of the robot, the lanes, the obstacles, and the robot’s control noise are modelled as non-parametric distributions as Gaussian assumptions of noise models are violated in real-world scenarios. Under these assumptions, we formulate a robust MPC that samples robotic controls effectively in a manner that aligns the robot to the goal state while avoiding obstacles and staying within the lane bounds under the duress of such non-parametric noise. In particular, the MPC incorporates a dis