About

William Saunders is a Research Engineer in the Alignment Team at OpenAI, where he works on developing AI systems that understand human values and behave in ways that align with them. His current work involves summarizing text with human feedback, as showcased in the blog post https://openai.com/blog/learning-to-summarize-with-human-feedback/. Prior to joining OpenAI, William worked as a Software Engineer at Ought, where he designed and implemented a web application for running experiments on factored cognition. He also collected a dataset of demonstrations of step-by-step answers to numerical estimation questions and implemented functionality for a Python forecasting library. Additionally, he worked as an intern at the Future of Humanity Institute, where he contributed to a paper on reinforcement learning and modified a Python/TensorFlow implementation of A3C to run reinforcement experiments on Atari games. William holds a Master of Mathematics in Computer Science (Human Computer Interaction) from the University of Waterloo and has completed additional graduate study in Machine Learning at the University of Toronto. He also has a Bachelor of Computer Science from the University of Waterloo and attended Waterloo Collegiate Institute for his high school education. William's technical skills include expertise in research science, AI, Python, software engineering, web development, reinforcement learning, and TensorFlow. He has 2.87 years of relevant experience in the field.