I am currently pursuing research at the Indian Institute of Science (IISc), where I’m a part of the Intelligent Inclusive Interaction Design Lab advised by Dr. Pradipta Biswas. Previously, I interned at Reliance Industries Limited as a Web Developer. Prior to that, I interned as a Machine Learning Engineer at Devtown. I recently completed my undergraduation (Bachelor of Engineering), in Information Technology from the G H Patel College of Engineering and Technology , Anand.
My long-term research goal is to develop autonomous systems which are capable of robust real-time decision-making. To achieve this, I've been working on deep learning for perception. Currently, I'm working on networks that can handle multiple perception tasks.
I believe the future of Advanced Driving Systems (ADS) perception lies in generic networks that can handle multiple perception tasks. Operational speeds higher than real-time ensure optimal response time for obstacles in autonomous driving systems. I want to develop models with an emphasis on interpretibility to tackle the current hinderences in reaching higher levels of autonomy.
Bandwidth Efficient Digital Image Watermarking Scheme Using a Concatenation of Three Transforms
International Conference on Interdisciplinary Research 2020
My projects range from lane and object detection on Indian roads to carpool systems at a refinery!
CGPA 8.95/10
91.0%
Proposed and designed a novel recurrent neural network for lane detection based on LSTM.
Achieved 2% improvement in accuracy and 150% reduction in inference cost on Indian Driving Dataset (IDD) over
CRF-based methods.
Integrate ROS object tracking, lane-line detection, and semantic segmentation architectures with AV software stack.
Spearhead development of Carpool Management System utilizing Node.js for the core trip execution engine, serving
24k users across a 300 sq. km area.
Reduced ETA by 30% using Dijkstra’s algorithm with heuristics, saving $560k in annual fuel costs.
Implemented a data visualization tool using react-map-gl to monitor 1480 vehicles.
Trained a YOLOv3 object detector to identify people and traffic signs on OpenImageV6 1.6M images, achieving a loss
of 4.8%.
Optimized a YOLOv3-Deep SORT object tracker through Dlib integration and minimized loss to 4.73%.
Led a team of eight students, conducted a session on Github Hacktoberfest, arranged logistics, and guided the team in organizing a Coding event for Imaze 2018 (National level Technical Symposium)
Planned delivery content and assisted two Gold Microsoft Learn Student Ambassadors from the University in delivering talks on Azure Cognitive Services at 5 State Colleges, engaging 650+ students.
Lead the Event Sponsorship and Social Media Marketing operations for a nationwide hackathon participated by 580 teams
Organised educational activities for underprivileged children.
Copyright © All rights reserved | This template is made with by Colorlib