Ideogram

Completed a computer science and engineering degree from the Indian Institute of Technology Bombay, then worked as a Research Assistant at Mila in Montreal before joining Google. At Google, progressed through roles including Software Engineer and Senior Research Scientist. As a researcher on Imagen at Google Brain, contributed to the foundational text-to-image diffusion model, serving as a lead author on the Imagen paper. Co-founded Ideogram in December 2022 with Mohammad Norouzi, William Chan, and Jonathan Ho, after the founders believed they could innovate faster outside a large corporate structure. This research-to-product transition directly shaped Ideogram's approach—specialized training data and architectural modifications from Saharia's expertise likely underpin the platform's exceptional text-rendering capabilities.

Studied electrical engineering and computer science at UC Berkeley (2010–2014) before pursuing a PhD in computer science at Berkeley (2017–2020) focused on unsupervised learning and reinforcement learning. Published Denoising Diffusion Probabilistic Models as lead author in 2020, a game-changing paper that reshaped the world of image generation. After his PhD, joined Google as a researcher working on diffusion models and helping develop Google's Imagen image generation model before co-founding Ideogram in 2022. His foundational work on diffusion-based generation informed Ideogram's core architecture for solving text-rendering precision in generated imagery.

Completed undergraduate studies at Sharif University of Technology in Iran, then earned a PhD in computer science from University of Toronto under the supervision of David Fleet. Joined Google Brain in January 2016, moving to the Toronto lab in 2018, where he eventually became Senior Staff Research Scientist. Led Imagen and received a NeurIPS 2022 Outstanding Paper Award for the influential text-to-image model. Left Google in December 2022 to co-found Ideogram, citing a desire to move faster outside large companies. Leadership approach emphasizes flat, cross-functional teams that "can move faster and build better and more coherent products than large hierarchical organizations."

Holds a PhD from Carnegie Mellon University and BASc from University of Waterloo, both in Computer Engineering. Served as Staff Research Scientist at Google Brain Toronto under Geoff Hinton, with prior internships at AMD, NVIDIA, Intel, Google, Amazon and TD Securities. At Google Brain, contributed to foundational research in text-to-image (Imagen, Imagen Editor), text-to-video (Imagen Video), and diffusion models across multiple modalities. Now Cofounder and CTO of Ideogram. Known for deep focus on systems-level optimization—built own distributed GPU training infrastructure from scratch in grad school and favors explicit control over hardware details (CUDA kernels, DMA transfers) and large-scale data infrastructure.