{"id":"allan-jabri","title":"Allan Jabri","content":"**Allan Jabri** is an artificial intelligence research scientist known for his work in self-supervised learning, computer vision, and reinforcement learning. He has held research positions at several major technology companies, including OpenAI, [Meta's Superintelligence Labs](https://iq.wiki/wiki/meta-superintelligence-team), and Facebook AI Research (FAIR). [\\[1\\]](#cite-id-anW2B4st8d) [\\[2\\]](#cite-id-eSvOvOstyx)\n\n$$widget0 [YOUTUBE@VID](https://youtube.com/watch?v=NO7Uo2ii1Sw)$$\n\n## Early Life\n\nJabri was born in Sydney, Australia, to a father of Lebanese descent and a mother of Chinese descent. He was raised primarily in the United States. According to the Paul & Daisy Soros Fellowships, his experience living between different cultures influenced his interest in understanding how simplicity can be distilled from complexity, which later catalyzed his curiosity in science and artificial intelligence. [\\[3\\]](#cite-id-TT29cTH4Tj)\n\n## Education\n\nJabri attended Princeton University, where he earned a Bachelor of Science (B.S.) in Computer Science in 2015. His undergraduate thesis focused on probabilistic methods for egocentric scene understanding and was awarded the Senior Thesis Prize in Computer Science. He later pursued his doctoral studies at the University of California, Berkeley, where he was a member of the Berkeley AI Research (BAIR) lab. He completed his Ph.D. in Computer Science between 2017 and 2023, advised by Professor Alexei A. Efros. His doctoral work was supported by the Paul & Daisy Soros Fellowship for New Americans, which he was awarded in 2018. [\\[3\\]](#cite-id-TT29cTH4Tj) [\\[4\\]](#cite-id-XiXtVDRUBR) [\\[6\\]](#cite-id-dS3wtUqhkP)\n\n## Career\n\nAfter graduating from Princeton, Jabri began his career as a Research Engineer at Facebook AI Research (FAIR) in New York. Following his time at FAIR, he commenced his Ph.D. at UC Berkeley. During his graduate studies, he undertook several research roles, including an internship at DeepMind in London and a position as a student researcher at Google Brain. Upon completing his Ph.D., Jabri joined OpenAI as a member of the technical staff and research scientist.\n\nIn July 2025, it was reported that Jabri, along with fellow OpenAI researcher [Lu Liu](https://iq.wiki/wiki/lu-liu), had been hired by [Meta Superintelligence Labs](https://iq.wiki/wiki/meta-superintelligence-team). This move was seen as part of Meta's broader strategy to bolster its artificial intelligence team and advance its capabilities in generative AI, following other high-profile hires and acquisitions in the field. The recruitment of top talent from competitors like OpenAI highlighted an industry-wide trend of major technology firms investing heavily to secure leadership in the expanding AI landscape. [\\[2\\]](#cite-id-eSvOvOstyx) [\\[1\\]](#cite-id-anW2B4st8d) [\\[5\\]](#cite-id-3nW0LcQcfx) [\\[7\\]](#cite-id-I6tfN4DqSR)\n\n## Research\n\nJabri's research primarily focuses on developing scalable objectives and architectures for self-supervised and unsupervised learning. His work often explores topics in continual learning, intrinsic motivation, and embodied cognition, with a long-term objective of creating learning algorithms that enable machines to autonomously acquire visual and sensorimotor common sense. He has contributed to numerous publications presented at major AI conferences such as NeurIPS, ICML, CVPR, and ICLR. [\\[1\\]](#cite-id-anW2B4st8d) [\\[3\\]](#cite-id-TT29cTH4Tj)\n\nHis notable works include:\n\n* **Space-Time Correspondence as a Contrastive Random Walk (NeurIPS 2020):** This paper introduced a method for learning dense visual representations from unlabeled videos by framing correspondence learning as a random walk on a space-time graph.\n\n$$widget0 [YOUTUBE@VID](https://youtube.com/watch?v=jzPGVHcQ87s)$$\n\n* **Learning Correspondence from the Cycle-Consistency of Time (CVPR 2019):** Co-authored with Xiaolong Wang and Alexei A. Efros, this research proposed using temporal cycle consistency in unlabeled videos to learn a generic representation for visual correspondence.\n* **Unsupervised Curricula for Visual Meta-Reinforcement Learning (NeurIPS 2019):** This work explored methods for the unsupervised discovery and meta-learning of visuomotor skills by applying deep clustering to an agent's own trajectories.\n* **Scalable Adaptive Computation for Iterative Generation (ICML 2023):** This research presented a universal neural architecture capable of adaptively allocating computation for the iterative generation of high-dimensional data, achieving strong results in image and video generation.\n\nA comprehensive list of his publications is available on his personal website and academic profiles. [\\[1\\]](#cite-id-anW2B4st8d) [\\[4\\]](#cite-id-XiXtVDRUBR) [\\[7\\]](#cite-id-I6tfN4DqSR) [\\[8\\]](#cite-id-4aTJPBiSMJ) [\\[9\\]](#cite-id-ylTZabR9zU)\n\n## Awards and Recognition\n\nIn 2018, Jabri was awarded the Paul & Daisy Soros Fellowship for New Americans. The fellowship supports immigrants and children of immigrants who are pursuing graduate studies in the United States. He also received the Senior Thesis Prize in Computer Science from Princeton University for his undergraduate work. [\\[3\\]](#cite-id-TT29cTH4Tj)","summary":"Allan Jabri is an AI research scientist known for his work in self-supervised and unsupervised learning. A Princeton and UC Berkeley alumnus, he has worked at OpenAI and Facebook AI Research (FAIR). In July 2025, Jabri joined Meta's AI team.","images":[{"id":"QmRJojyjyq5ioi3hg82n34ejQaqhf2S341cq5LBq6yQcCq","type":"image/jpeg, image/png"}],"categories":[{"id":"people","title":"people"}],"tags":[{"id":"AI"},{"id":"Developers"},{"id":"PeopleInDeFi"}],"media":[{"id":"QmYL7kZdzLUPyv3EUBnLZpFoF1AggXYXff3MfrazAUpUo5","name":"norway3.png","caption":"","thumbnail":"QmYL7kZdzLUPyv3EUBnLZpFoF1AggXYXff3MfrazAUpUo5","source":"IPFS_IMG"},{"id":"QmXfvAeAyxvsdjXKk3kT5SatCsXGnoKV9i9xfZJHf5ikHa","name":"1448621.jpeg","caption":"","thumbnail":"QmXfvAeAyxvsdjXKk3kT5SatCsXGnoKV9i9xfZJHf5ikHa","source":"IPFS_IMG"},{"id":"https://www.youtube.com/watch?v=NO7Uo2ii1Sw","name":"NO7Uo2ii1Sw","caption":"","thumbnail":"https://www.youtube.com/watch?v=NO7Uo2ii1Sw","source":"YOUTUBE"},{"id":"https://www.youtube.com/watch?v=jzPGVHcQ87s","name":"jzPGVHcQ87s","caption":"","thumbnail":"https://www.youtube.com/watch?v=jzPGVHcQ87s","source":"YOUTUBE"}],"metadata":[{"id":"references","value":"[{\"id\":\"anW2B4st8d\",\"url\":\"https://ajabri.github.io/\",\"description\":\"Personal website\",\"timestamp\":1756112130062},{\"id\":\"eSvOvOstyx\",\"url\":\"https://www.ainvest.com/news/meta-hires-openai-researchers-2507/\",\"description\":\"Meta Hires Two OpenAI Researchers\",\"timestamp\":1756112130062},{\"id\":\"TT29cTH4Tj\",\"url\":\"https://pdsoros.org/fellows/allan-jabri/\",\"description\":\"Allan Jabri – Paul & Daisy Soros Fellowships\",\"timestamp\":1756112130062},{\"id\":\"XiXtVDRUBR\",\"url\":\"https://openreview.net/profile?id=~Allan\\\\_Jabri2\",\"description\":\"OpenReview Profile for Allan Jabri\",\"timestamp\":1756112130062},{\"id\":\"3nW0LcQcfx\",\"url\":\"https://x.com/ajabri\",\"description\":\"A Jabri on X\",\"timestamp\":1756112130062},{\"id\":\"dS3wtUqhkP\",\"description\":\"LinkedIn: Allan Jabri\",\"timestamp\":1756112216530,\"url\":\"https://www.linkedin.com/in/allanjabri/\"},{\"id\":\"I6tfN4DqSR\",\"description\":\"Google Scholar: Allan Jabri\\n\",\"timestamp\":1756112573037,\"url\":\"https://scholar.google.com/citations?user=_QlCijoAAAAJ\"},{\"id\":\"4aTJPBiSMJ\",\"description\":\"Learning Visually Grounded Sentence Representations\",\"timestamp\":1756112623703,\"url\":\"https://aclanthology.org/people/allan-jabri/\"},{\"id\":\"ylTZabR9zU\",\"description\":\"Alyosha Efros & Allan Jabri - Space-Time Correspondence as a Contrastive Random Walk\\n\",\"timestamp\":1756112836318,\"url\":\"https://www.youtube.com/watch?v=jzPGVHcQ87s\"}]"},{"id":"website","value":"https://ajabri.github.io/"},{"id":"linkedin_profile","value":"https://www.linkedin.com/in/allanjabri/"},{"id":"github_profile","value":"https://github.com/ajabri"},{"id":"twitter_profile","value":"https://x.com/ajabri"},{"id":"email_url","value":"mailto:ajabri@gmail.com"},{"id":"previous_cid","value":"\"https://ipfs.everipedia.org/ipfs/QmSnw7fgNDo6WbbSLxGkrtr4T5oNcArKtxaQFBGEBJhWPi\""},{"id":"commit-message","value":"\"docs: Enrich Allan Jabri article with embedded YouTube videos.\""},{"id":"previous_cid","value":"QmSnw7fgNDo6WbbSLxGkrtr4T5oNcArKtxaQFBGEBJhWPi"}],"events":[{"id":"e6060924-ff49-4f1f-bcee-7ed55d40184e","date":"2015-05","title":"Graduated from Princeton University","type":"DEFAULT","description":"Graduated from Princeton University with a Bachelor of Science in Computer Science. His thesis work on egocentric scene understanding won the Senior Thesis Prize.","multiDateStart":null,"multiDateEnd":null},{"id":"f8bae04b-b571-4c67-a860-051e56fa4857","date":"2018-04","title":"Awarded PD Soros Fellowship","type":"DEFAULT","description":"Received the Paul & Daisy Soros Fellowship for New Americans to support his PhD studies in Computer Science at the University of California, Berkeley.","multiDateStart":null,"multiDateEnd":null},{"id":"969103e0-c7de-44c7-aced-d15d69d60dfc","date":"2023-05","title":"Completed PhD at UC Berkeley","type":"DEFAULT","description":"Completed his PhD in Computer Science at UC Berkeley, where he was a member of the Berkeley AI Research (BAIR) lab, advised by Alexei A. Efros.","multiDateStart":null,"multiDateEnd":null},{"id":"f234745a-1f02-4706-b3bd-8ac163450011","date":"2025-07","title":"Joined Meta's AI Team","type":"DEFAULT","description":"Was hired by Meta Platforms to join its artificial intelligence (AI) team, moving from his previous role as a research scientist at OpenAI.","multiDateStart":null,"multiDateEnd":null}],"user":{"id":"0x8af7a19a26d8fbc48defb35aefb15ec8c407f889"},"author":{"id":"0x8af7a19a26d8fbc48defb35aefb15ec8c407f889"},"language":"en","version":1,"linkedWikis":{"blockchains":[],"founders":[],"speakers":[]}}