Gaurav Bombe

- Developing and deploying advanced detection algorithms for edge devices, with a focus on enhancing camera-based applications - Training detection models, optimizing them to run efficiently on Neural Processing Units (NPUs), and implementing object tracking and counting functionalities - Maximize the use of limited hardware resources on edge devices to deliver client-ready applications. - Translating client requirements into actionable processes and developing proof of concepts.

-Researching the effect of self-supervised pre-training for learning image representations on object detection and metric learning (k-NN classification) tasks. -Used contrastive learning methods: Swapping Assignments between Views (SwAV), SEER - FAIR 2020 and DetCon - DeepMind 2021 -Research on the effect of data augmentation schemes on learning image representations.

-Developed construction process optimization algorithms and computer vision solutions. -Developed algorithms to track construction developments like concreting, workers detection using object detection (YOLOv5). -Optimized existing analyzing algorithms. Utilized image aggregation to remove camera overlaps, customizing the video data timelapse

Core Subjects: Deep Learning, Computer Vision, Software Engineering, Control Systems, Automation & Industrial Communication, C++ Programming, Sensorics.