AI People Wiki

Builders & open source

Yuval Alaluf

Research Scientist, Runway

Yuval Alaluf is a Research Scientist at Runway, working on video synthesis and creative applications including projects such as Runway Aleph, an in-context video model for editing, transforming, and generating video. Alaluf was previously a Research Scientist at Pika focusing on video synthesis and an intern at Snap Research. Alaluf's PhD research at Tel Aviv University, advised by Daniel Cohen-Or, explored how large-scale generative models can be controlled, edited, and personalized. Alaluf is first author of ReStyle, a StyleGAN encoder, and co-author of Attend-and-Excite, a guidance method for text-to-image diffusion models; both official repositories are MIT-licensed.123456

At a glance

Current role
Research Scientist
Company
Runway
Sector
Company
Work listed
2
Sources cited
12
Profile reviewed

Career

Alaluf earned a PhD in computer science at Tel Aviv University, advised by Daniel Cohen-Or, with research on how large generative models can be controlled, edited and personalized. Earlier, Alaluf led applied research projects in few-shot learning, object detection, representation learning and uncertainty estimation.1

Alaluf interned at Snap Research, exploring personalization in generative models, and was a research scientist at Pika working on video synthesis. At Runway, Alaluf works on video synthesis and creative applications, including Runway Aleph, Aleph 2.0 and Solaris.12

Research

Alaluf's early papers concerned StyleGAN. With Or Patashnik and Daniel Cohen-Or, Alaluf worked on encoders that map real images into StyleGAN's latent space for editing, including e4e (2021), ReStyle (2021), which refines an inversion iteratively, and HyperStyle (2021), which uses hypernetworks to adjust the generator for each image.7389

Alaluf then turned to text-to-image diffusion models. Alaluf co-authored Textual Inversion (2022), which personalizes a model by learning a new word embedding from a few images, and Attend-and-Excite (2023), which guides cross-attention so that every subject in a prompt appears. Later work includes MyVLM (2024) for personalizing vision-language models and Solaris (2026), a model that generates an interactive interface frame by frame in response to user actions.1051112

Work and contributions

  • Runway Aleph Proprietary product

    Alaluf's homepage names Runway Aleph among the video synthesis projects Alaluf works on at Runway; Runway describes Aleph as an in-context video model for editing, transforming, and generating video.12

  • ReStyle and Attend-and-Excite Open-source code

    First author, with Or Patashnik and Daniel Cohen-Or, of the 2021 ReStyle paper on a StyleGAN encoder that works by iterative refinement, whose MIT-licensed code is published under Alaluf's GitHub account, and second author of "Attend-and-Excite: Attention-Based Semantic Guidance for Text-to-Image Diffusion Models" (2023), whose official MIT-licensed repository is also hosted there.34569

Timeline

  1. Co-authored "Designing an Encoder for StyleGAN Image Manipulation" (e4e).7
  2. Posted ReStyle, a StyleGAN encoder based on iterative refinement, as first author.3
  3. Posted HyperStyle, StyleGAN inversion with hypernetworks, as co-first author.8
  4. Co-authored Textual Inversion for personalizing text-to-image generation.10
  5. Co-authored Attend-and-Excite as co-first author.59
  6. Posted MyVLM, on personalizing vision-language models for user-specific queries, as first author.11
  7. Runway introduced Aleph, an in-context video model that Alaluf worked on.21
  8. Posted Solaris, an interface world model, as first author.12

Other documented roles

Roles the sources document that are not confirmed as current.

  • Research Scientist, Pika1

Where they work

Suggest a correction or removal

Profiles cover public professional work only, and each statement cites a source. If something is out of date, missing or wrong, tell us and link a source. People can also ask for their profile to be corrected or removed.

References

12 sources · profile reviewed Oct 21, 2026 · data updated Oct 21, 2026

  1. Yuval Alaluf — homepageYuval Alaluf · Official page · accessed Oct 21, 2026
  2. Introducing Runway Aleph — Runway ResearchRunway · Announcement · published Jul 25, 2025 · accessed Oct 21, 2026
  3. ReStyle: A Residual-Based StyleGAN Encoder via Iterative Refinement (arXiv 2104.02699)arXiv · Paper · published Apr 6, 2021 · accessed Oct 21, 2026
  4. yuval-alaluf/restyle-encoder LICENSE (MIT)yuval-alaluf (GitHub repository) · License · accessed Oct 21, 2026
  5. Attend-and-Excite: Attention-Based Semantic Guidance for Text-to-Image Diffusion Models (arXiv 2301.13826)arXiv · Paper · published Jan 31, 2023 · accessed Oct 21, 2026
  6. yuval-alaluf/Attend-and-Excite LICENSE (MIT)yuval-alaluf (GitHub repository) · License · accessed Oct 21, 2026
  7. Designing an Encoder for StyleGAN Image Manipulation (arXiv 2102.02766)arXiv · Paper · published Feb 4, 2021 · accessed Oct 21, 2026
  8. HyperStyle: StyleGAN Inversion with HyperNetworks for Real Image Editing (arXiv 2111.15666)arXiv · Paper · published Nov 30, 2021 · accessed Oct 21, 2026
  9. Yuval Alaluf — researchYuval Alaluf · Official page · accessed Oct 21, 2026
  10. An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversion (arXiv 2208.01618)arXiv · Paper · published Aug 2, 2022 · accessed Oct 21, 2026
  11. MyVLM: Personalizing VLMs for User-Specific Queries (arXiv 2403.14599)arXiv · Paper · published Mar 21, 2024 · accessed Oct 21, 2026
  12. Solaris: Towards Interfaces That Are Generated, Not Coded (arXiv 2609.00776)arXiv · Paper · published Sep 1, 2026 · accessed Oct 21, 2026