AI People Wiki

Builders & open source

Matthew Zeiler

Clarifai founder named SVP, Research at Nebius in May 2026; author of the ADADELTA paper

Matthew Zeiler founded Clarifai, and Nebius announced in May 2026 that Zeiler, then Clarifai's founder and CEO, would join Nebius as SVP, Research, to head a team working on multimodal agentic reasoning, world models, token efficiency, and long-term memory, alongside Clarifai's core engineering and research team. With Rob Fergus at New York University's Courant Institute, Zeiler co-authored the 2013 paper Visualizing and Understanding Convolutional Networks. Zeiler is also the sole author of the 2012 ADADELTA paper on per-dimension adaptive learning rates, written during an internship at Google.123

At a glance

Current role
None confirmed in the sources
Sector
Company
Education
NYU
Work listed
2
Sources cited
4
Profile reviewed

Research

Zeiler's research papers came from work with Rob Fergus at New York University's Courant Institute and an internship at Google. In 2012 Zeiler was sole author of ADADELTA, a per-dimension learning-rate method for gradient descent that needs no manual tuning of a global learning rate, and in early 2013 Zeiler and Fergus proposed stochastic pooling, which regularizes convolutional networks by sampling activations within each pooling region.34

"Visualizing and Understanding Convolutional Networks" (2013), with Fergus, introduced a technique for visualizing what intermediate layers of a large ImageNet-trained convolutional network respond to, and used it to diagnose and improve the network's architecture.2

Career

Zeiler founded Clarifai and led it as founder and CEO. In May 2026 Nebius announced that Clarifai's core engineering and research team, led by Zeiler, was joining Nebius, and that Nebius was licensing Clarifai's inference and compute orchestration technology for its Nebius Token Factory platform. The announcement said Zeiler would join Nebius as SVP, Research, leading a team on multimodal agentic reasoning, world models, token efficiency and long-term memory.1

Work and contributions

Timeline

  1. Posted "ADADELTA: An Adaptive Learning Rate Method" as sole author, based on work during a Google internship.3
  2. Posted "Stochastic Pooling for Regularization of Deep Convolutional Neural Networks" with Rob Fergus.4
  3. Posted "Visualizing and Understanding Convolutional Networks" with Rob Fergus.2
  4. Nebius announced that Zeiler would join as SVP, Research, along with Clarifai's core team.1

Documented roles

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

  • Founder and CEO (as described in Nebius's May 2026 announcement), Clarifai1
  • Intern (the ADADELTA work was done during the internship), Google3
  • Listed with the Department of Computer Science, Courant Institute, on the 2013 paper Visualizing and Understanding Convolutional Networks, New York University2

Companies in their history

  • Google 121 people in the wiki
  • NYU 12 people in the wiki

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 21, 2026 · data updated Oct 21, 2026

  1. Nebius welcomes Clarifai's core team and licenses inference IP to strengthen Nebius Token FactoryNebius · Announcement · published May 12, 2026 · accessed Oct 21, 2026
  2. Visualizing and Understanding Convolutional Networks (arXiv 1311.2901)arXiv · Paper · published Nov 12, 2013 · accessed Oct 21, 2026
  3. ADADELTA: An Adaptive Learning Rate Method (arXiv 1212.5701)arXiv · Paper · published Dec 22, 2012 · accessed Oct 21, 2026
  4. Stochastic Pooling for Regularization of Deep Convolutional Neural Networks (arXiv 1301.3557)arXiv · Paper · published Jan 16, 2013 · accessed Oct 21, 2026