Summary
James Mcinerney is a Senior Research Scientist in New York with 13 years of experience applying probabilistic machine learning and generative AI to large-scale personalization and recommendation systems. At Netflix he drives methodological advances in epistemic uncertainty, variational inference for temporal point processes, and ML-driven simulation for long-term user satisfaction, building on prior leadership roles at Spotify where he pioneered bandits in production and counterfactual evaluation for slate recommendations. His work bridges deep theory and production: papers at NeurIPS, KDD, RecSys and WSDM reflect both new inference methods (e.g., the Implicit Delta Method) and practical evaluation tools that inform product launches. Trained at Southampton, Imperial and Oxford, he combines an academic rigor from postdocs at Columbia and Princeton with hands-on ML4Sys engineering, and often focuses on making complex Bayesian ideas operational at scale. Notably, he contributes research that explicitly quantifies uncertainty to improve decision-making in content personalization—a less visible but critical lever for safer, more effective recommender systems.
13 years of coding experience
5 years of employment as a software developer
MA Computer Science, MA Computer Science at University of Oxford
Doctor of Philosophy (PhD), Artificial Intelligence, Doctor of Philosophy (PhD), Artificial Intelligence at University of Southampton
MSc Artificial Intelligence, MSc Artificial Intelligence at Imperial College London