I am a Senior Computer Vision Engineer working on perception related problems at Innovasea Systems.
Before joining Innovasea, I was a robotics engineering Master's student at the University of Maryland, College Park. My interests lie at the intersection of the fields of vision and lidar-based perception for Robotics. At UMD, I was a Research Assistant in the Robotics Algorithms and Autonomous Systems Lab at the University of Maryland, advised by Dr. Pratap Tokekar. I worked on implementing robust perception pipeline for the precise mapping of bridges from UAVs.
My work in the industry and in academia includes an overlap of : Visual Odometry, Object Segmentation, Object Detection and Tracking, Visual SLAM, Structure from Motion (SfM).I did my undergrad in Instrumentation and Control Engineering at Nirma University, Ahmedabad where I worked with Prof. Sandip Mehta and Dr. Dilip Kothari.
Computer Vision Engineer
Innovasea Systems (January 2021 - current)
My current job here involves designing and improving underwater perception solution for accurate biomass estimation of fish using underwater perception sensors.
Software Engineer, Computer Vision
Vecna Robotics, Boston (August 2020 - December 2020)
Implemented perception pipeline for robustly detecting pose of payloads from a forklift robot equipped with a lidar and a camera. Improved detection mAP scores by 9% than existing models and deployed mechanism to train models at scale with metrics logging.
Perception Intern
Aeva Inc., Mountain View (May 2019 - August 2019)
Quantifying object tracking and detection in the perception pipeline of Aeva Lidar package. Implemented and integrated end to end metrics to set benchmarks for tracker and classifier improvements.Research Assistant
RAAS Lab, University of Maryland (September 2019 - August 2020)
I am currently a Research Assistant in the Robotics Algorithms and Autonomous Systems Lab at the University of Maryland, advised by Dr. Pratap Tokekar. Currently working on deploying perception pipeline for UAV based bridge infrastructure inspection. The sensor module includes a 3D Lidar (VLP-16), ZED Camera and IMU.
Research Assistant
Computer Vision Lab, University of Maryland (August 2018 - May 2019)
Implementation of supervised deep learning models for optical flow for use on edge inference devices such as Intel Neural Compute Stick.