About

Ajay Prasadh Viswanathan is a Data Scientist at Microsoft, specializing in machine learning for systems. He has a Master's degree in Artificial Intelligence from the University of Edinburgh, where he focused on building character-level neural networks for text classification. With over 5 years of relevant experience, Ajay has been studying, prototyping, building, and deploying ML systems in language processing and computer vision. At Microsoft, Ajay is part of the Office365 infrastructure team, where he builds machine learning systems to predict resource consumption per user for load balancing scenarios, build accurate time aggregators, and ensure optimal utilization of hardware in data-centers. Previously, he was an AI resident with Microsoft Research, where he worked with the Hololens team to use deep learning for novel view synthesis in computer vision. He built a pipeline to do a 3D geometric transform and used deep learning models (UNet, Conditional GAN) to generate novel views. Before joining Microsoft, Ajay worked as a Creative Data Scientist at To Play For, where he brought characters to life from books. He extracted dialogue utterances and built a language model of a character, designed, developed, and deployed intent classification, emotion detection, and sentence ranking systems. He also developed a semi-supervised machine learning algorithm to reduce false positives in text classification. Ajay's tech stack includes Research Scientist, Data Scientist, ML, AI, Computer Vision, Text Classification, Deep Learning, Neural Networks, Infra, GAN, and language models. He is excited about the future of computing and AI and is looking forward to playing a part in defining our computing experiences.