About

Vishal Kumar is an experienced data scientist with a strong background in Natural Language Processing (NLP) models, classical machine learning algorithms, and computer vision. He has a Bachelor of Technology from the prestigious Indian Institute of Technology, Bombay, and is skilled in Python programming language, statistical data analysis, machine learning, and deep learning. Currently, Vishal is the Lead Assistant Manager of the Digital R&D Team at EXL, where he analyzes the business process of insurance companies to reduce their recurring costs. He implements digital transformation using ML & NLP solutions and restructures the business process. Vishal has performed exploratory data analysis to gain insights from the data using Python and built text classification models for few entities using Logistics Regression, Random Forest, SVM, and BERT to achieve accuracy of approximately 90%. He has also developed custom SpaCy NER to extract structured output with an accuracy of 80% extraction efficiency and improved the logic for prediction of one entity, resulting in a 20% increase in efficiency by doing RCA. Vishal has deployed the NER and classification models on AWS cloud using GitLab and AWS lambda handlers. He has helped the client save about 80% time in their process, resulting in expected $50 million savings. In the past, Vishal has worked as an Assistant Manager of the Digital R&D Team at EXL and as a Computer Vision Intern at Artivatic.ai. At Artivatic.ai, he incorporated image pre-processing techniques using OpenCV libraries in Python to improve text extraction accuracy in Tesseract. He built a parser for key-value mapping of structured text extracted from images into a dictionary by using RegEx, Fuzzywuzzy, and achieved an accuracy of 90.2% on matching the actual requirement with respect to automated model output. Vishal has a relevant experience of 2.28 years and a strong tech stack that includes NLP, AWS, BERT, ML, Research Scientist, Data Scientist, Text Classification, Computer Vision, Python, Deep Learning, SageMaker, SQL, ResNet, SVM, XGBoost, Random Forest, RNN, CNN, Word2Vec, TensorFlow, Object detection, Keras, OpenCV, Object Detection, and Scikit.