About

Harsh Sharma is a passionate Data Scientist with 5.78 years of relevant experience in the field. Currently working at FIS as a Data Scientist, he has previously worked at Akaike Technologies as a Data Scientist and TCS as a Software Developer. During his time at Akaike Technologies, Harsh worked on a range of computer vision projects, including Face Recognition, where he used Retinaface based on Resnet100 architecture to detect faces in a given frame and using ArcFace based on Resnet50 to generate vectors for the detected faces. He also optimized the whole network for that to run on embedded devices with low specifications. Another notable project was the Automated Inspection of Defects in Turbines, where he used segmentation model built on resnet34 and Unet architecture to locate the edges of turbines in a frame using segmentation and make vision-based automated inspection of the defects in gas turbines. Additionally, he worked on Counting Unique Footfalls in a Showroom using video footage. At TCS, Harsh worked as a Python developer, building a scalable pipeline for ETL process within a secure environment and speeding the process up. He also worked in Innovation Labs to enhance his skills in machine learning, doing various projects including Four Wheeler Demand Prediction, Cancer Type Prediction, Dominant Color Extraction, Style Transfer, and more. Harsh's education background includes a B.Tech in ECE from Guru Gobind Singh Indraprastha University, which he completed in 2017. His tech stack includes expertise in Data Science, Research Science, Computer Vision, ML, Deep Learning, AI, Python, Pytorch, UI, Azure, Software Engineering, ResNet, Web, Seaborn, Sagemaker, OpenCV, Image Segmentation, NLP, Scikit, Redis, Tensorflow, Pandas, MySQL, DevOps, and GCP. Harsh is passionate about making machines intelligent and is always looking for new challenges to enhance his skills and knowledge in the field.