Jackson Stokes

Jackson Stokes is a Cofounder at TrainLoop, a Y Combinator-backed artificial intelligence infrastructure startup developing a platform for the reasoning fine-tuning of large language models. His current company builds specialized tools that allow software developers to enhance complex AI systems through advanced data curation, reward model training, and streamlined inference deployment. Prior to launching TrainLoop, Stokes worked as an engineer at Google, where he focused on deep learning infrastructure and executed critical kernel-level optimizations for the flagship Gemini generative AI model. During his tenure, he also served at Google Research, conducting framework-level optimizations on advanced video models to improve overall computational efficiency and system performance. Stokes leverages his extensive engineering experience to build highly efficient machine learning frameworks for enterprise applications. His current focus remains on making sophisticated reinforcement learning techniques from major AI laboratories accessible to developers.