About

Kushal Chauhan is a Research Software Engineer at Google Research India, with a focus on developing machine learning methods that are both robust and interpretable. He is particularly interested in leveraging principled probabilistic methods to achieve these goals. With nearly four years of relevant experience, Kushal has worked on improving the robustness and explainability of deep learning models across unsupervised, supervised, and reinforcement learning settings. His work includes unsupervised out-of-distribution detection, collaborative prediction systems, and hierarchical reinforcement learning. Kushal has published several papers on his research, including a long paper accepted at CVPR 2022 and a long paper accepted at AAAI 2023. In addition to his work in AI, Kushal has a keen interest in reading about Physics, Math, and Philosophy. He frequently visits blogs like Shtetl-Optimized, Terrence Tao's blog, and math3ma to stay up-to-date on the latest developments in these fields. Kushal is also an avid viewer of educational YouTube channels like 3Blue1Brown, PBS Space Time, and Veritasium. Kushal holds an Integrated Post Graduate degree in Information Technology from ABV-Indian Institute of Information Technology and Management. He also has a Higher School Secondary Certificate (XII) and a Secondary School Certificate (X) from Lalaji Memorial Omega International School. Prior to his current role at Google, Kushal worked as a Pre-Doctoral Researcher at the company, where he was mentored by Research Scientist Pradeep Shenoy. He also worked as a Data Science Intern at ShareChat, where he worked on creative text generation using transformers and GANs to assist in content creation for creators on the platform. Kushal's knowledge and experience in the field of AI, as well as his curiosity and passion for learning, make him a valuable asset to any team.