PeopleofSuperintelligence

Researcher

Yoad Tewel

Yoad Tewel is a computer vision researcher whose publications span zero-shot image captioning, object grounding and text-to-image generation. He coauthored Perfusion, which personalizes generated images, and ConsiStory, which keeps subjects consistent across prompts. His 2025 papers introduced methods to control object counts and insert objects into images from text instructions, and in 2026 he coauthored a framework for unpaired image and video editing. As of October 2026, he is a research scientist at NVIDIA Research and a Computer Science PhD candidate at Tel-Aviv University under Professor Lior Wolf.

Now at
NVIDIA

Timeline

  1. 2022

    Zero-shot image captioning

    In 2022, he coauthors ZeroCap, a zero-shot method that combines a visual-semantic model with a large language model to generate descriptive text from an image without extra training or tuning.

  2. 2022

    Open-world object grounding

    In 2022, he coauthors a weakly supervised, open-world phrase-grounding method that returns object boxes and descriptive phrases from an image without text input.

  3. 2023

    Perfusion for personalization

    In 2023, he coauthors Perfusion, a text-to-image personalization method that uses dynamic rank-1 updates to adapt a model to personalized concepts.

  4. 2024

    Consistent text-to-image generation

    In 2024, he coauthors ConsiStory, a training-free method that shares a pretrained model's internal activations to keep subjects consistent across images while preserving prompt alignment.

  5. 2025

    Counting objects with prompts

    In 2025, he coauthors Make It Count, whose CountGen method generates the number of objects specified in a prompt while keeping a natural layout.

  6. 2025

    Training-free object insertion

    In 2025, he coauthors Add-it, a training-free method that extends diffusion models' attention to add objects to images from text instructions.

  7. 2026

    Bootstrap Your Generator

    In 2026, he coauthors Bootstrap Your Generator, a framework for unpaired image and video editing that uses a pretrained model's knowledge without paired data.

  8. As of October 2026

    Research at NVIDIA and Tel-Aviv

    He is a research scientist at NVIDIA Research and a Computer Science PhD candidate at Tel-Aviv University, working in the Deep Learning Lab under Prof. Lior Wolf.

Career

  1. NowNVIDIAResearch Scientist

Education

  1. Tel Aviv UniversityPhD

From the biography.

  1. Google has the highest combined h-index among AI organizationsIncludes NVIDIA. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.
  2. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes NVIDIA. We estimate that 539 of the 6,766 people with a known current organization (8%) hold both a university role and a company role, including 20% of people at companies and 13% of people at universities.
  3. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes NVIDIA. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.
  4. Universities (31%) and startups and other companies (25%) employ the largest shares of 2020–2026 entrantsIncludes NVIDIA. Among 819 people whose first recorded role started in 2020–2026 and who have a known current organization, 60% work in commercial sectors (25% at startups and other companies, 23% at frontier labs and 13% at big tech, with 9 in two), and 31% work at universities, compared with 66% for earlier entrants.

Sources and credits

Photo: We have no published photo of Yoad Tewel yet. Send us a free-licensed photo.

  1. ZeroCap: Zero-Shot Image-to-Text Generation for Visual-Semantic Arithmetic openaccess.thecvf.com. Checked 20 October 2026.
  2. What is Where by Looking: Weakly-Supervised Open-World Phrase-Grounding without Text Inputs proceedings.neurips.cc. Checked 20 October 2026.
  3. Key-Locked Rank One Editing for Text-to-Image Personalization arxiv.org. Checked 20 October 2026.
  4. Training-Free Consistent Text-to-Image Generation arxiv.org. Checked 20 October 2026.
  5. Make It Count: Text-to-Image Generation with an Accurate Number of Objects make-it-count-paper.github.io. Checked 20 October 2026.
  6. Add-It: Training-Free Object Insertion in Images with Pretrained Diffusion Models mlanthology.org. Checked 20 October 2026.
  7. Bootstrap Your Generator: Unpaired Visual Editing with Flow Matching arxiv.org. Checked 20 October 2026.
  8. Yoad Tewel — Research Scientist, NVIDIA Research yoadtew.github.io. Checked 20 October 2026.