About

Gaurav Kumar is a highly skilled Senior Data Scientist with over 6 years of experience in developing AI solutions utilizing cutting-edge developments in the field of data science. Currently working at Juniper Networks, Gaurav is responsible for implementing state-of-the-art Deep Learning solutions. He is also the brain behind the development of an open-source library, Fastestimator, which integrates Tensorflow and PyTorch in a single unified framework to simplify deep-learning model development for GE Healthcare AI eco-system. Gaurav's expertise lies in data augmentation, model compression, and reinforcement learning. He has worked on developing approaches for neural network compression by conducting experiments on classification, segmentation, and detection tasks. During his time at General Electric, Gaurav worked on the development of a deep learning framework integrating Tensorflow and PyTorch for GE Healthcare AI eco-system. He also worked on productization of AI models to reduce memory footprint and improve training and inference speed. Prior to his current role, Gaurav worked as a Consultant at Absolutdata Analytics, where he designed a beverage assortment optimization engine based on machine learning/deep learning techniques for one of the world's largest beverage companies. He also designed data-driven marketing campaigns to increase revenue and improve customer engagement for one of Malaysia's leading Oil and Gas companies. Gaurav holds a Bachelor of Technology (B.Tech.) in Electronics and Electrical Communication Engineering from IIT Kharagpur. He has a strong technological background and is proficient in various programming languages such as Python, SQL, and AWS. He is also experienced in using tools and packages such as Scikit, Keras, XGBoost, Random Forest, Seaborn, Sagemaker, Apache Hive, and Pandas. In summary, Gaurav Kumar is a highly skilled and experienced Senior Data Scientist who has contributed significantly to the development of AI solutions in the field of data science. His expertise in data augmentation, model compression, and reinforcement learning has been invaluable in the development of state-of-the-art Deep Learning solutions.