Avatar

Ben London

Principal Scientist

Amazon Music

About Me

I am a Principal Scientist at Amazon Music. Broadly speaking, I research machine learning theory and algorithms, using theoretical analysis to inform the design of new ML algorithms.

I earned my Ph.D. in 2015 at University of Maryland, where I was advised by Lise Getoor and worked closely with Ben Taskar and Bert Huang. My dissertation studied generalization in structured learning, and its relationship to the algorithmic stability of collective inference. I have also worked on: generalization guarantees for randomized learning (such as stochastic gradient methods); recommendation/personalization systems; contextual multi-armed bandits; and counterfactual learning from logged bandit feedback.

Interests

  • Statistical Learning Theory
  • Structured Prediction and Probabilistic Graphical Models
  • Bandits and Learning from Logged Bandit Feedback
  • Recommendation and Personalization

Education

  • PhD in Computer Science, 2015

    University of Maryland

  • MSc in Computer Science, 2008

    Columbia University

  • BMu in Music Technology, 2001

    New York University

Recent Posts

Industry Co-chair for RecSys 2024

I am honored to serve as Industry Co-chair for RecSys 2024.

Four new papers at RecSys 2023 workshops

I recently presented four new papers at the RecSys 2023 workshops

AC for ICLR 2024

I will be an area chair for ICLR 2024.

WebConf 2023 Tutorial on ''Practical Bandits''

I helped create a tutorial on ‘‘Practical Bandits’’ at The Web Conference 2023

New paper accepted at AISTATS 2023

My recent paper on boosting for off-policy learning has been accepted at AI & Statistics (AISTATS) 2023.

Experience

 
 
 
 
 

Principal Scientist

Amazon

Apr 2023 – Present Seattle, WA
I mainly work on contextual bandits and offline (off-policy) evaluation and learning, for developing music recommendation systems.
 
 
 
 
 

Sr. Scientist

Amazon

Oct 2018 – Mar 2023 Seattle, WA
I mainly work on contextual bandits and offline (off-policy) evaluation and learning, for developing music recommendation systems.
 
 
 
 
 

ML Scientist

Amazon

Sep 2015 – Sep 2018 Seattle, WA
I started my career at Amazon on the Core ML team, working on semantic similarity for product recommendation. I then moved to Amazon Music’s ML team, where I continued to work on recommendation.
 
 
 
 
 

Research Intern

Google Research

May 2014 – Aug 2014 New York, NY
Explored structured prediction methods for WebTables.
 
 
 
 
 

Graduate Research Assistant

University of Maryland, LINQS Lab

Jan 2011 – Sep 2015 College Park, MD
Lise Getoor’s research lab (now located at UC Santa Cruz).
 
 
 
 
 

Analytic Software Engineer

Sentrana

Sep 2008 – Aug 2010 Washington, DC
Productionized statistical models.
 
 
 
 
 

Software Engineer

Anthrotronix

Feb 2008 – Aug 2010 Siver Spring, MD