Prafulla Dhariwal
Also known as prafdhar
Prafulla Dhariwal is an OpenAI researcher whose generative-model work spans music, images and language. He won olympiad gold medals in astronomy, mathematics and physics, and then studied computer science at MIT. He joined OpenAI as a research intern in 2016 and became a research scientist in 2017. In 2020 he co-wrote Jukebox, a music model, and the GPT-3 paper, which NeurIPS named a 2020 Best Paper. He leads the Omni team that builds GPT-4o, OpenAI's first natively fully multimodal model.
Timeline
-
2009–2013
Gold medals at three olympiads
He wins gold medals at the International Astronomy Olympiad in 2009, the International Mathematical Olympiad in 2012 and the International Physics Olympiad in 2013.
-
2013
Studying computer science at MIT
In 2013 he takes the IIT-JEE exam, then joins MIT in the US, where he studies computer science and mathematics.
-
2016–2017
Joining OpenAI as an intern
He joins OpenAI as a research intern in 2016 and becomes a research scientist in 2017.
-
April 2020
Jukebox, a music generator
In April 2020 he is one of six authors of Jukebox, a model that generates music with singing.
-
2020
A NeurIPS 2020 Best Paper
In 2020 the GPT-3 paper, Language Models are Few-Shot Learners, wins a NeurIPS 2020 Best Paper Award, and he is one of its authors.
-
May 2021
Diffusion models beat GANs
He co-writes Diffusion Models Beat GANs on Image Synthesis, a paper posted to arXiv in May 2021.
-
Since 2024
The Omni team at OpenAI
He leads the Omni team at OpenAI, which builds GPT-4o, the company's first natively fully multimodal model, launched in May 2024.
Sources and credits
Photo: We have no published photo of Prafulla Dhariwal yet. Send us a free-licensed photo.
- OpenAI CEO Sam Altman praised Prafulla Dhariwal - The Global Indian
- Sam Altman Credits Pune Boy Prafulla Dhariwal For Creation Of New GPT-4o Model
- Prafulla Dhariwal: The mind behind OpenAI’s GPT-4o
- [2005.00341] Jukebox: A Generative Model for Music
- Announcing the NeurIPS 2020 award recipients – NeurIPS Blog
- [2005.14165] Language Models are Few-Shot Learners
- [2105.05233] Diffusion Models Beat GANs on Image Synthesis
- Meet Prafulla Dhariwal, the Indian American Prodigy Behind GPT-4o, OpenAI’s Latest Innovation - American Kahani