About

Rajat Agarwal is an Applied Scientist II at Amazon Ads, where he is responsible for building large-scale machine learning systems for Traffic Quality, specifically Robot and Ad Fraud Detection. He is passionate about building ML-driven products and is interested in deep learning applied to computer vision, NLP, semi and self-supervised learning, generative modeling, and model-free reinforcement learning. Currently, Rajat is a Research Intern at the Applied Sciences and Machine Learning Group at Microsoft Research India. Here, he has worked on deep reinforcement learning and contextual bandits for downstream metric optimization and its applications in software engineering, parameter tuning, and neural architecture search. Rajat has previously worked as an Applied Scientist II at Amazon, where he focused on building deep learning models for ad relevance and ranking, keyword matching, click prediction, and ad auction bid optimization. He has also worked as a Software Development Engineer Intern at Amazon, where he developed convolutional neural networks for attribute extraction, semantic segmentation, and large-scale similarity search on images. Rajat holds a degree in Computer Science and Chemistry from Birla Institute of Technology and Science, Pilani. He has 4.86 years of relevant experience and is skilled in research science, reinforcement learning, machine learning, deep learning, supervised learning, computer vision, NLP, search, software engineering, and neural networks.