At Cognizant, Jason performs groundbreaking research that combines evolutionary computation and deep learning to create powerful new algorithms for revolutionizing the field of artificial intelligence. His current areas of research encompass neural architecture search, multi-task learning, multi-objective optimization, computer vision, and transfer learning. He has experience with multiple programming languages and frameworks, including Tensorflow, Keras, Python, C++, shell scripting, and AWS. His long term goals are to develop a technological foundation for making artificial general intelligence (AGI) possible and to philosophize about the impact of AGI in relation to modernity.
Prior to that, he interned at Sentient Technologies (later acquired by Cognizant) where he leveraged massively distributed compute to scale up neural architecture search and achieve state of the art results in multiple benchmarks and domains. Jason received a Bachelor’s degree from the University of California, Berkeley and a Master of Science and PhD degrees from the University of Texas at Austin.
His academic research includes applying bilevel optimization to complex control tasks and developing multi-agent systems for game playing. He also conducted research on evolutionary methods for evolving deep neural network architectures, and invented a novel algorithm that uses coevolution to discover structures commonly seen in state of the art, hand designed networks. This work has resulted in several patents and publications at academic conferences.