Badges
Certifications
Work Experience
Associate Engineer
Mando Corporation•  April 2020 - Present
Pedestrian Detection algorithm development for Automated Parking System - Developed a novel deep learning-based object detection model having a model size of ~5MB and running at ~100FPS (on 2080Ti) which is compatible with automotive grade production boards, having accuracy comparable to the algorithms like YOLOv3, SSD, etc. [Patent pending] - Working on the improvements of the developed deep-learning model to meet the high standards of detection accuracy required in life-critical systems of ADAS. - Working on post-training quantization and the problems associated with it.
Graduate Engineer
Mando Corporation•  July 2019 - April 2020
Perception algorithm development for Advanced Driver Assistance System (ADAS) - Worked on the development of a novel deep-learning based object detector capable of running at real-time speeds on automotive grade boards which have a limited compute and memory availability. - Contributed to the development of a road marking detection model which will aid the functionality of lane keep assist feature of ADAS. - Contributed to post-training model quantization which is essential for deployment of any deep-learning model on an edge device. - Worked with huge amounts (several terabytes) of unstructured data. - Took up the responsibility of establishing and maintaining the in-house GPU server.
Trainee Engineer
Mando Corporation•  September 2018 - July 2019
- Grasped knowledge of the state-of-the-art in various deep learning-based vision tasks. - Understood the problem of perception algorithms and the difficulties associated with the development of such algorithms required in Advanced Driver Assistance Systems like adaptive cruise control, lane keep assist, automated parking, etc.
Education
GL Bajaj Institute of Technology & Management
Computer Science & Engineering, B.Tech•  2015 - 2019