Aditya Ramesh
Aditya Ramesh is an OpenAIOpenAIOpenAI is an artificial intelligence research and product organization founded in 2015 that develops the GPT model series, ChatGPT, Codex and a developer API, and conducts research on image, speech, video and intelligent agents. After its 2025 restructuring, the nonprofit OpenAI Foundation continues to control the for-profit OpenAI Group PBC. Its growth has involved large-scale fundraising and compute partnerships, and has also been accompanied by controversies over governance, copyright, privacy and AI safety.Open the full entry researcher who created the text-to-image model DALL·E and co-led the Sora video-generation effort. After graduating from New York UniversityNew York UniversityNew York University is a private university in New York City, founded in 1831, and one of the largest private universities in the United States.Open the full entry, he turned an OpenAI internship into a research career and went on to lead Worldsim, exploring how video models could learn about the physical world.
Contents10 sections
Key facts
Student years at NYU and internship at OpenAI
Ramesh did his undergraduate work at New York University and took part in research in Yann LeCun’s lab. He considered applying for a doctorate, but decided to stay at OpenAI after a summer internship. In a 2024 conversation with NYU’s Saining Xie at the Beijing Academy of Artificial Intelligence conference, he revisited that decision and discussed opportunities for researchers in evaluation and interpretability.[1]
2021–2023From DALL·E to conversational image generation
In January 2021, OpenAI introduced DALL·E, which Ramesh had developed to combine concepts described in words into a single image. DALL·E 2, which he co-created, added local edits and image variations in April 2022. In 2023 his team brought the third model to ChatGPTChatGPTChatGPT is a generative AI assistant launched by OpenAI on November 30, 2022. Through websites and applications, it provides conversation, writing, coding, search, file analysis, and image and voice interaction. It evolved from a free research preview into personal, enterprise and education plans, gradually adding deep research, Study mode and cross-application task execution. Its use has prompted continuing discussions about factual accuracy, personal privacy, copyright, education and teen safety.Open the full entry, allowing people to refine image descriptions in conversation. On the second anniversary of the first release, he said in an interview that the speed with which text-to-image systems became tools for the public had surprised him.[2][3][4]
April 25, 2024Interview with No Priors
Ramesh appeared on No PriorsNo PriorsNo Priors is an AI interview podcast and media brand founded and hosted by investors Sarah Guo and Elad Gil, with episodes beginning in February 2023. It covers research, startups, infrastructure and business applications through guest interviews, host discussions and retrospectives distributed on audio platforms and YouTube.Open the full entry with Tim BrooksTim BrooksTim Brooks is an AI researcher specializing in image and video generation. At the University of California, Berkeley, he developed InstructPix2Pix for editing photographs with written instructions. He went on to co-lead Sora at OpenAI and work on multimodal generation at Google DeepMind and Meta Superintelligence Labs.Open the full entry and Bill PeeblesBill PeeblesBill Peebles is an artificial intelligence researcher, first author of the diffusion transformer (DiT) paper, who led OpenAI’s video-generation model Sora and, after leaving OpenAI in 2026, co-founded an AI video startup for filmmakers.Open the full entry to explain which kinds of motion, geometry and temporal relationships Sora could simulate and where it still failed. They also discussed artists’ need for control over visual style and the safety evaluation preceding a wider release.[5]
He recalled artists wanting to give a model their portfolios so that later image requests could draw on their own earlier work.[5]
2024-07: Midyear retrospective
The July midyear retrospective replayed the team's discussion of artists working with Sora.[6]
2024-12: Year-end retrospective
The year-end anthology revisited the Sora team's discussion of visual models and artificial general intelligence at 11:17.[7]
From 2024Video generation and robotics research
In his account of the December 2024 release, Ramesh said he assembled the team that built both Sora's model and its website experience. He described his contributions to Sora 2Sora (OpenAI)Sora is a text-to-video model and app of the same name developed by OpenAI; it debuted as a research preview in February 2024, was followed by Sora 2 and a social video app in 2025, and its web and app services were discontinued in April 2026.Open the full entry in 2025 as improving the connection between text and generated video and giving users greater control. As a vice president of research leading OpenAI's Worldsim effort, he subsequently extended this work to robots' understanding of the physical world.[3][1]
2026Robotics team and reorganization
On May 31, 2026, Sam AltmanSam AltmanSam Altman is an entrepreneur and investor, and the co-founder and chief executive officer of OpenAI. He previously served as president of Y Combinator and founded the location-based social service Loopt. He also helped found Tools for Humanity and Merge Labs, has invested in energy and life sciences ventures, and has driven public discussion of AI governance, employment, and wealth distribution.Open the full entry identified Ramesh as the leader of the world-simulation effort that had developed into OpenAI Robotics over the preceding year. On November 13, OpenAI confirmed to Fortune that it had reorganized the stand-alone robotics group, integrating research and applied staff into larger teams while retaining robotics as a long-term research priority.[8][9]
Advice Shared on No Priors
He urged distinguishing the responsibilities of model providers, distribution platforms and creators, including how viewers might learn whether a source was trustworthy. (20:45)[5]
He suggested developing tools that reflected creators' own aesthetics, including the design vocabulary particular to a studio. (16:26)[5]
Related Organizations, Websites and Public Accounts
OpenAI: His biography lists him as a vice president of research who led Worldsim. Official website: https://openai.com/ (opens in a new window).[3]
Sources9 sources
- 谢赛宁与Sora负责人最新对话实录:靠互联网视频就够训练出AGI吗? (opens in a new window) (in Chinese)thepaper.cn↑
- Two years after DALL-E debut, its inventor is “surprised” by impact (opens in a new window)venturebeat.com↑
- Aditya Ramesh (opens in a new window)adityaramesh.com↑
- DALL·E 3 (opens in a new window)openai.com↑
- No Priors Ep.61 | OpenAI's Sora Leaders Aditya Ramesh, Tim Brooks and Bill Peebles (opens in a new window)youtube.com↑
- No Priors Ep. 71: The Best of 2024 (so far) with Sarah Guo and Elad Gil (opens in a new window)youtube.com↑
- No Priors Ep. 95 | Best of 2024 (opens in a new window)youtube.com↑
- OpenAI Robotics is hiring (opens in a new window)x.com↑
- Exclusive: OpenAI disbands stand-alone robotics team as general AI models start running robots (opens in a new window)fortune.com↑