Jonathan Valverde

Machine Learning Software Engineer

Google DeepMind

About me

I am a Machine Learning Software Engineer at Google DeepMind performing research on recommender systems.

I graduated with an M.S from the Department of Computer Science at the University of Maryland, College Park (UMD), where I conducted research on recommender systems advised by Prof. John Dickerson. I completed my undergraduate education at Princeton University, where I majored in Mechanical and Aerospace Engineering, with minors in Computer Science and Robotics.

At UMD, I was funded by a GEM Fellowship with Amazon as my GEM employer sponsor. Through this fellowship, I completed two internships as an Applied Scientist Intern at Amazon Prime Video, where I worked on research projects aimed at improving personalized video recommendations.

I am broadly interested in applied research in machine learning, recommender systems, and multi-armed bandits.

All views expressed are my own.

Download my resumé.

Interests

Machine learning
Recommender systems
Multi-armed bandits

Education

MS in Computer Science, 2022

University of Maryland, College Park (UMD)
BSE in Mechanical Engineering, Minor in Computer Science, 2016

Princeton University

Sample Projects

Multi-armed bandits for targeted advertising in social networks

We posed the problem of maximizing user engagement with advertisements in a social network, where highly influential users are prioritized. This framework was inspired by social networks in which a few users, often called influencers, are followed by a large number of users (for instance, in Twitter and Instagram). In our proposed setting, we must use observations from any related followers to better select ads for an influencer.

Senior thesis: multi-armed bandits for robotic field estimation task

The problem of field estimation, or finding the spatial distribution of a resource in space, has many applications in problems of robotic search. In my thesis, I approached the problem in the framework of a Gaussian Multi-Armed Bandit (MAB) task. The arms correspond to discretized points in space, and the smoothness of the field is modeled as spatial correlation between the arms.