Lex Beattie
I research algorithmic impact at Spotify and hold a Ph.D. from the University of Oklahoma, where I focused on the evaluation and mitigation of algorithmic bias in recommendation systems.
Spotify
2016 – Current
I have worn many hats during my time at Spotify. Over the past seven years, I have worked as a data scientist, data engineer, ml infrastructure engineer, and researcher.
I manage a team of data and research scientists focusing on quantifying, evaluating, and mitigating algorithmic impact in music and podcast recommendation systems at Spotify. Before working in algorithmic impact, I founded Spotify’s ML Engagement team with a mission to democratize and improve the state of ML across Spotify. With that team, I helped over 50 different teams across Spotify understand ML best practices, productionize ML workflows and implement impactful ML in their products.
University of Oklahoma
Master’s of Science, Data Science & Analytics
Ph.D., Engineering
My journey in academia started at McGill University with a Bachelor’s of Commerce in International Management. After developing my talent for data working in the industry, I decided to take my technical skills to the next level with a Master’s of Science in Data Science & Analytics. I rediscovered my love of learning and research with the University of Oklahoma.
Upon graduating, I decided to continue my academic career and pursue a Ph.D. in Engineering specializing in Machine Learning. My research focused on fairness, bias, and harm in recommendation systems. I believe in creating safe and trustable AI systems, my work seeks to lower the barrier in implementing responsible recommendation systems in practice. It is becoming paramount that ML practitioners understand how their systems work and their potential to harm downstream users. The goal of my research to increase the transparency of complex ML systems and create trust with their users.
-
Evaluation Framework for Understanding Sensitive Attribute Association Bias in Latent Factor Recommendation Algorithms. Lex Beattie, Isabel Corpus, Lucy H. Lin, Praveen Ravichandran. arXiv preprint arXiv:2310.20061 (2023)
Scoping Fairness Objectives and Identifying Fairness Metrics for Recommender Systems: The Practitioners’ Perspective. Jessie J. Smith, Lex Beattie, Henriette Cramer. In Proceedings of the ACM Web Conference 2023 (WWW ‘23)
Challenges in Translating Research to Practice for Evaluating Fairness and Bias in Recommendation Systems. Lex Beattie, Dan Taber, Henriette Cramer. In Sixteenth ACM Conference on Recommender Systems (RecSys ’22).
RecSys Fairness Metrics: Many to Use But Which One to Choose? Jessie J. Smith, Lex Beattie. In FAccTRec Workshop at Sixteenth ACM Conference on Recommender Systems (RecSys '22).
Building Human Values into Recommender Systems: An Interdisciplinary Synthesis. Jonathan Stray, Alon Halevy, Parisa Assar, Dylan Hadfield-Menell, Craig Boutilier, Amar Ashar, Lex Beattie et al. Planned publication in ACM Transactions on Recommender Systems.
A machine learning and clustering-based approach for county-level COVID-19 analysis. Charles Nicholson, Lex Beattie, Matthew Beattie, Talayeh Razzaghi, and Sixia Chen. Plos one 17, no. 4 (2022): e0267558.
-
Data Science in 2021 - AnacondaCon 2021
From Idea to Production ML - MLOps.community 2021
Team Teardown: Driving Platform Adoption and Success at Spotify - TWIMLcon 2021
Building High-Performance ML Teams - apply(conf) 2021