Responsibilites

• Taught Math, Programming and Deep Learning architectures using Python, PyTorch and many ML libraries.

• Coordinated and guided more than 40 students to work on AI projects from scratch.

• Performed weekly hands-on webinars on vital concepts of AI from math to implementing a deep learning model.

• Aided the students with a 1:1 call whenever needed to streamline a personalized learning experience.

Responsibilites

• Built a Sportsperson Tracking System to estimate and track efficient behavior of a player on the field using Deep Learning (Pose Estimation) in TensorFlow and standard OpenCV methods.

• Implemented a Mask-RCNN based segmentation and feature extraction deep learning application for applying apparel to human bodies in a virtual cloth try-on system.

• Customized implementation of Image Classification, Object Detection, Custom key point detection using CNN, Image Segmentation frameworks in TensorFlow and Keras for application in sports and fashion.

• Adapted a standard GAN architecture to train and optimize for a customized virtual clothing experience which can generate an image of given apparel on a person.

• Worked on License Plate Recognition, Person Counting & Tracking, Photogrammetry and 3D reconstruction.

Responsibilites

• Architected and Developed algorithms, software for a simultaneously operated multi-robot warehouse system.

• Collaborated with the hardware engineer to integrate and interface with the robot on several fronts.

• Designed a virtual grid mapping system for the warehouse and streamlined a collision avoidance path planning algorithm for the autonomous robots using A*.

• Introduced a Centralized Processing System which communicates with the stateless robots using MQTT.

• Overhauled an ARUCO marker-based localization system (Raspberry Pi) with a RFID based localization system (NodeMCU), effectively cutting down the cost, hardware and space requirements of the robot.