Mohammad Norouzi
Mohammad Norouzi is a machine learning researcher and AI lab leader whose work spans self-supervised learning, generative audio and text-to-image models. After undergraduate studies at Sharif University of Technology in Iran, he completed a computer science PhD at the University of Toronto under David Fleet in 2015. At Google Brain, he coauthored SimCLR and WaveGrad in 2020 and Imagen in 2022, whose paper received a NeurIPS Outstanding Paper award. He left Google to cofound Ideogram, which raised seed and Series A funding and released its 3.0, 4.0 and 4.5 models; as of October 2026, he is its cofounder and CEO.
Timeline
-
By 2015
Sharif and a Toronto PhD
From Iran, he finishes his undergraduate studies at Sharif University of Technology and moves to Canada at age 22 for graduate school. He completes a computer science PhD at the University of Toronto in December 2015 under David Fleet; a Google PhD fellowship supports his research on scalable similarity search.
-
2016–2018
Joining Google Brain
He joins the Google Brain team in Mountain View in January 2016 and moves to Toronto in January 2018, where he works as a senior staff research scientist.
-
2020
SimCLR and WaveGrad
In 2020, he coauthors SimCLR with Ting Chen, Simon Kornblith and Geoffrey Hinton. A linear classifier trained on its self-supervised representations reaches 76.5% top-1 accuracy on ImageNet. He also coauthors WaveGrad, a waveform-generation model built on prior work on score matching and diffusion probabilistic models, with Nanxin Chen, Yu Zhang, Heiga Zen, Ron Weiss and William Chan.
-
2022
Imagen and a NeurIPS award
He coauthors Imagen, a text-to-image diffusion model that uses large language models such as T5 to encode the prompt. The paper receives a NeurIPS 2022 Outstanding Paper award.
-
2022–2023
Co-founding Ideogram
In 2022, he and William Chan, Chitwan Saharia and Jonathan Ho cofound Ideogram, a Toronto startup that builds its own text-to-image model from scratch. The four founders are Imagen researchers at Google Brain; by September 2023, Norouzi is the company’s chief executive as it announces US$16.5 million in seed funding led by a16z and Index Ventures.
-
2024
Ideogram 2.0 and Series A
In February 2024, Ideogram raises $80 million after its 1.0 model release. In June 2024, Norouzi describes his move from researcher to startup CEO on the a16z AI podcast, and the company releases its 2.0 model in August 2024.
-
March 2025
Ideogram 3.0
In March 2025, Ideogram releases its 3.0 model with improved realism and a better understanding of complex text layout.
-
As of October 2026
Ideogram cofounder and CEO
In June 2026, Ideogram releases its open-weight 4.0 model, and Norouzi discusses the company’s open-weight decision with a16z partners Yoko Li and Justine Moore. The company releases 4.5 in September 2026. As of October 2026, he is Ideogram’s cofounder and CEO.
Career
- Since 2022IdeogramCofounder and CEO, CEO
- 2016–2022Google BrainSenior Staff Research Scientist
Education
Sources and credits
Photo: We have no published photo of Mohammad Norouzi yet. Send us a free-licensed photo.
- Mohammad Norouzi
- Toronto startup Ideogram launches image generator as interest in AI tools shows signs of slowdown - The Globe and Mail
- A Simple Framework for Contrastive Learning of Visual Representations
- WaveGrad: Estimating Gradients for Waveform Generation
- Photorealistic Text-to-Image Diffusion Models with Deep Language Understanding
- NeurIPS 2022 Awards
- Ideogram (text-to-image model) - Wikipedia
- The Future of Image Models Is Multimodal | Andreessen Horowitz
- AI, Design, and the Power of Open Models | Andreessen Horowitz
- Ideogram: org chart, alumni mafia and talent flow | Crustdata Research






