AI People Wiki

Builders & open source

Hassan Shapourian

First author of Zyphra's Zamba2-VL technical report and ZAYA1-VL-8B release on open vision-language models

Hassan Shapourian is the first author of the Zamba2-VL technical report (2026), which describes Zyphra's open vision-language models built on the Zamba2 hybrid SSM-transformer backbone at 1.2B, 2.7B and 7B parameters and released under Apache 2.0. Shapourian is also the first-listed author of Zyphra's May 2026 ZAYA1-VL-8B release, and in 2024 was credited, with a Cisco affiliation, as a co-author of Zyphra's post on pruning the deeper layers of language models.1234

At a glance

Current role
None confirmed in the sources
Work listed
2
Sources cited
4
Profile reviewed

Research

Shapourian's Zamba2-VL report, posted in May 2026 with Kasra Hejazi, Olabode M. Sule and Beren Millidge, adapts Zyphra's Zamba2 hybrid state-space language models to image understanding. Zyphra's release post describes the family as competitive with transformer-based open vision-language models of similar size while having lower time-to-first-token.12

The same author group is credited on Zyphra's ZAYA1-VL-8B release from May 2026. Earlier, in March 2024, Shapourian was listed with a Cisco affiliation among the co-authors of Zyphra's post "The Unreasonable Ineffectiveness of the Deeper Layers".34

Work and contributions

  • Zamba2-VL-7B Open weights

    First author of the Zamba2-VL technical report, which presents the Zamba2-VL family of vision-language models at 1.2B, 2.7B and 7B parameters.12

  • ZAYA1-VL-8B Open weights

    First-listed author of Zyphra's ZAYA1-VL-8B release post on an open visual-language model.3

Timeline

  1. Listed, with a Cisco affiliation, as a co-author of Zyphra's post on pruning deeper layers.4
  2. First-listed author of Zyphra's ZAYA1-VL-8B vision-language model release.3
  3. Posted the Zamba2-VL technical report on arXiv as first author.1
  4. Zyphra released the Zamba2-VL models with Shapourian as first-listed author.2

Documented roles

The sources do not confirm a current role; these are the roles they document.

  • Credited first-listed author on Zyphra vision-language model releases (2026), Zyphra23
  • Affiliation listed on a 2024 Zyphra research post, Cisco4

Companies in their history

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 If something is out of date, missing or wrong, tell us and link a source.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

4 sources · profile reviewed Oct 22, 2026 · data updated Oct 22, 2026

  1. Zamba2-VL Technical Report (arXiv 2606.00390)arXiv · Paper · published May 29, 2026 · accessed Oct 22, 2026
  2. Zamba2-VL: Hybrid SSM Vision-Language Models — ZyphraZyphra · Announcement · published Jun 2, 2026 · accessed Oct 22, 2026
  3. ZAYA1-VL-8B: Efficient Open Visual Intelligence — ZyphraZyphra · Announcement · published May 8, 2026 · accessed Oct 22, 2026
  4. The Unreasonable Ineffectiveness of the Deeper Layers — ZyphraZyphra · Official page · published Mar 28, 2024 · accessed Oct 22, 2026