About

Dhruv Premi is a highly skilled Senior AI Researcher II at American Express AI Labs with over 9 years of relevant experience in the field of artificial intelligence. He is a life-long learner and is constantly seeking to expand his knowledge in the areas of AI, research, and development. Dhruv is passionate about using his knowledge and skills to create socially helpful products that can make a difference in people's lives. He hopes to become a part-time teacher, engineer, player, novelist, and musician one day. Prior to his current role, Dhruv has worked as a Data Science Manager at Citi Research and Development Lab, where he was the owner of Form OCR. He implemented a cursive handwritten recognition system using a deep learning network comprising of CNN, LSTM, and CTC layers, achieving state-of-the-art accuracy. He also improved the accuracy of the speech recognition system by including the capability of multiple language models in the prefix beam search phase. Dhruv has also worked as a Data Scientist/Machine Learning Expert at Wipro HOLMES (AI Platform) product wing of Wipro, where he filed a patent for a system and method for text to speech conversion of an electronic document. He has worked on several products for WIPRO LTD, including EKYC and Contract Management, and multiple POCs. Dhruv is the owner of Text Alignment Framework, which detects tables from PDF pages, segments a PDF page into individual paragraphs, and uses a novel method that utilizes General CRF. He is also the owner of Sequence Labeling Framework, which streamlines the process of manual creation of datasets, and part-owner of NLP Framework, which has the flexibility to choose multiple libraries and ingestion of new features for the named entity extraction part. Dhruv holds a Master's degree in Computer Science from the Indian Institute of Technology, Kharagpur, and a Self-Driving Car Nano Degree from Udacity. He has a strong technical skill set, including expertise in AI, research scientist, deep learning, data science, CNN, language models, ML, NLP, image processing, TensorFlow, and search.