Saurabh Koju is a Computer Engineering student and Research Assistant at NAAMII working in the MultiModal Learning Lab. His research interests span multimodal learning, reinforcement learning, and fundamental machine learning theory.
He is particularly interested in understanding the theoretical foundations that enable machine learning systems to work effectively, as well as exploring how different data modalities can be combined to create more capable AI systems. His work in reinforcement learning focuses on developing agents that can learn optimal behaviors through interaction with their environment.