I work at TellApart, where I develop models to help e-commerce sites display ads to those customers that react most positively to them. TellApart is tackling challenging machine learning problems at web-scale and has lots of opportunities to contribute. Please reach out to me if you're interested in hearing more!
My research interests revolve around machine learning and data at scale, specifically modeling incentivized behavior. More general research interests include statistical machine learning, reinforcement learning, cognitive science, and cognitive modeling.
My dissertation investigates learning agents that have deliberate control over internal explicit memories. As an analogy, humans learn to use working memory, episodic memory, and semantic memory in order to perform in the world. Similarly, the learning agents in my dissertation use biologically-inspired memory models in order to learn to perform in their environments; furthermore, they learn to use these memories while simultaneously performing in their environments. The dynamics of learning both over memory and over task actions gives rise to interesting phenomena which are identified and investigated. My research unifies the foundational statistical reinforcement learning algorithms of SARSA and Q-learning with biologically-inspired memory models of cognitive science.
