Experience

Oct 2021 – Oct 2023

AI Researcher

Honda R&D Co. Ltd., Honda Innovation Lab, Tokyo

  • Advanced Driver Assistance System: built a low-latency, low-cost smart DashCam and presented the proof of concept to CXOs and the board.
  • Cooperative Micro-Mobility: built a robust pedestrian trajectory prediction framework and a dynamic simulation environment.
  • Contributed to path planning and robot vision (object detection, monocular depth) and wrote performance-optimised C++ for the real-time system. Acted as the main debugger for system stability and latency.
  • Participated in ICRA 2022: reviewed papers and attended workshops.
  • Received a special performance bonus in 2 out of 3 performance reviews.

Mar 2021 – Jun 2021

NLP Software Development Intern

LimeChat, India

  • Built a tree-based product quiz for the chatbot and added support for multiple variants of the same product.
  • Designed a data-management framework to visualise training and improve intent and entity classification.

May 2020 – Jul 2020

Research Intern

Alberta Machine Intelligence Institute (Amii), Edmonton, Canada

  • Studied the effect of state similarities on generalisation in Monte-Carlo Tree Search using the AlphaZero architecture.
  • Built a bot on top of KataGo, the state-of-the-art Go bot, using Memory-Augmented MCTS.

Nov 2019 – Dec 2019

Research Intern

RIKEN Center for Advanced Intelligence Project, Tokyo

  • Built a reinforcement learning framework for robots using Deep Q-Learning and MCTS, integrated with ROS.
  • Designed a car-racing toy problem in ROS where the bot navigates a maze optimally.

Sep 2018 – Jan 2020

Core Team Member

Robomuse 5.0, IIT Delhi

  • Deployed stereo vision algorithms to extract depth data for SLAM on a mobile robot.
  • Developed the complete back end of a fully autonomous navigating robot using ROS navigation packages.
  • Designed a vision-based landmark recognition technique to estimate the robot’s global position.
  • Developed human and object detection modules to add safety features.

May 2019 – Jul 2019

Research Intern

RIKEN Center for Advanced Intelligence Project, Tokyo

  • Studied traffic state prediction during the Hiroshima disaster of June 2018.
  • Compared RNNs with conventional machine learning methods (SVMs, random forests).
  • Related the spatio-temporal features learned by the models to traffic-flow theory. This work led to a publication.

Feb 2018 – Aug 2018

Project Associate

D-LIVE, Mahindra Autonomous Car Challenge, IIT Delhi

  • Improved odometry by 90% at 35 km/h by fusing a ZED camera, GPS and IMU, with correct loop closure.
  • Worked on ORB-SLAM2 for camera trajectory and sparse 3D reconstruction.
  • Received the Design Innovation Summer Award (DISA) and a Letter of Recommendation.
Download Resume