Mikhail Trofimov is an AI researcher and engineer at OpenAI focused on AI for health.https://mtrofimov.com/https://mtrofimov.com/blog/beneficial-rl/https://mtrofimov.com/blog/beneficial-rl/Can reinforcement learning on realistic beneficial behavior make models more broadly and persistently aligned?Mon, 22 Jun 2026 00:00:00 GMThttps://mtrofimov.com/projects/beneficial-rl/https://mtrofimov.com/projects/beneficial-rl/Reinforcement learning research on alignment that generalizes across domains and persists under adversarial pressure.Mon, 22 Jun 2026 00:00:00 GMThttps://mtrofimov.com/blog/healthbench-professional/https://mtrofimov.com/blog/healthbench-professional/A physician-designed benchmark for evaluating language models on realistic and challenging clinical conversations.Thu, 30 Apr 2026 00:00:00 GMThttps://mtrofimov.com/projects/healthbench-professional/https://mtrofimov.com/projects/healthbench-professional/A physician-designed benchmark for evaluating language models on realistic clinical conversations.Thu, 30 Apr 2026 00:00:00 GMThttps://mtrofimov.com/projects/tesla-fsd/https://mtrofimov.com/projects/tesla-fsd/Computer-vision research and model post-training for production autonomous-driving systems.Sat, 01 Oct 2022 00:00:00 GMThttps://mtrofimov.com/projects/recommender-systems/https://mtrofimov.com/projects/recommender-systems/Research on joint representation learning and pairwise ranking for implicit-feedback recommendation.Mon, 01 Mar 2021 00:00:00 GMThttps://mtrofimov.com/projects/tffm-exponential-machines/https://mtrofimov.com/projects/tffm-exponential-machines/Open-source and research work on efficient arbitrary-order feature interactions using tensor methods.Thu, 12 May 2016 00:00:00 GMThttps://mtrofimov.com/projects/kaggle-competitions/https://mtrofimov.com/projects/kaggle-competitions/Competition machine learning across malware classification, search relevance, product classification, and science QA.Tue, 01 Jul 2014 00:00:00 GMT