About

Nimish Nahar is a highly skilled Data Scientist with 1.28 years of relevant experience in the field. Currently working at claim genius as a Data Scientist, Nimish has made significant contributions to the field of Computer Vision and Deep Learning. He has constructed an orientation correction model that corrects the orientation of given car/partial car images for feeding them to later Deep Learning models in the pipeline. Nimish has generated generic orientation correction data for car images due to lack of standard datasets and performed data augmentations for better learning. He has achieved classification accuracy of 99.12% for a 4 class model, i.e., 0, 90, 180 and 270 degrees with VGG-16 backbone. Nimish has extended the work by modifying the model to detect the exact angle by which the image was disorientated, i.e., any angle between 0-359. He has defined a custom metric to evaluate the performance aligned with the goal, which also suited the downstream tasks. Nimish has used ResNet50v2 and EfficientNetB2 as backbone networks and achieved an average angle error of 2-3 degrees. He has deployed the model and served as an API using the flask framework. Prior to his current role, Nimish worked as a Data Scientist Trainee at claim genius where he gained hands-on experience in the field of Computer Vision and Deep Learning. He has also worked as a Project Intern at Virtusa, where he researched, prototyped and applied state-of-the-art technology (BERT) to build an Open Domain Question Answering system. Tensor-flow version of the well-known NLP model, i.e., BERT, which was made available by Hugging Face, was used as the basis of the model. Nimish holds a Bachelor of Technology in Electrical and Electronics Engineering from the Indian Institute of Technology, Roorkee. He has a strong technical skill set, including Data Scientist, Research Scientist, NLP, Computer Vision, Python, Deep Learning, ML, Flask, BERT, TensorFlow, OpenCV, ResNet, Software Engineer, Image Processing, Neural Networks, Search, and Pytorch. Nimish is a highly motivated individual who is passionate about solving complex problems and delivering high-quality results. He is a quick learner and has excellent communication and interpersonal skills.