CHINMAYANANDA SAHU

Worked on a 6DoF pose estimation pipeline for small objects using object detection from YOLOv8, SAM for segmentation, Mobilenetv3 as classifier, pose estimation from megapose and template pose for robotic arm pickup. Implemented a custom lightweight model for face detection and age and gender classification, achieving a notable 94% accuracy on the Asian faces dataset for edge deployment. Developed of a real time automated Deep Learning algorithm-based LULC classifier system team using satellite images (Sentinel-2 & Landsat-8 images).

Responsible for executing all the phases of machine learning life cycle including data collection, Preparation, Data Processing, Data Analysis, Implementation of Machine Learning Algorithm, Creating Deep Learning Architecture. Developed and Implemented multi class object detection model and increase its accuracy to 85% for camera. Tested models we trained on the Sony’s intelligent sensor for accuracy and edge device compatibility of view. Worked on classifying emotions such as happiness, sadness, anger, surprise, fear, and disgust. By analyzing key facial features like eye movements, eyebrow position, lip curvature, and overall facial expressions with 90% accuracy.

Volunteered in Ashmaan Foundation as an Psychology Intern in 2020

Relevant coursework: Statistics, R Programming, Python Programming, Machine Learning, Artificial Intelligence, Data Engineer

Relevant coursework: Physics, Mathematics, Quantum Computing, Quantum Optics