Nicholas Saban

Member of Technical Staff, Research

Nick is a member of technical staff at Patronus AI and a coauthor on the FigmaTrace dataset and the GlobeBench benchmark. He studies data science at UC Berkeley and coauthored AutoAdv, a framework for automated multi-turn red teaming of language models, presented at a NeurIPS 2025 workshop.

Google Scholar ↗LinkedIn ↗X ↗

At Patronus· Newest first

Newest first

Papers

“Ganges River Delta”, Bay of Bengal“Ganges River Delta”, Bay of Bengal

GlobeBench

A benchmark for whether text world models keep track of state, rules and side effects.

Nicholas Saban, Charvi Bannur, Mariya Vasileva, Darshan Deshpande
arXiv 2610.22708
NeurIPS 2026 Workshop
October 15, 2026
Workshop
globebench
https://arxiv.org/abs/2610.22708
NeurIPS 2026 Workshop
“Melted Colors”, Great Sandy Desert, Western Australia“Melted Colors”, Great Sandy Desert, Western Australia

FigmaTrace

Human design workflows in Figma, captured as training data for creative design tasks.

Darshan Deshpande, Yoshinari Fujinuma, Martyna Markiewicz, Devanshu Bansal, Shivani Jain, Nicholas Saban, Chirag Maheshwari, Anand Kannappan
arXiv 2608.21460
August 20, 2026
figmatrace
https://arxiv.org/abs/2608.21460

Posts

Corona Arch, Moab, Utah. Photo: Vulturesong (CC0)

Introducing FigmaTrace: A Comprehensive Training Dataset for Figma Design Workflows

Aug 20, 2026
·
Research
figmatrace-a-comprehensive-training-dataset-for-figma-design-workflows
Darshan Deshpande
Nicholas Saban
Anand Kannappan
Rain over the tallgrass prairie, Tallgrass Prairie National Preserve, Kansas

No papers or posts here yet.

Work with Nick

See open roles →
Nicholas Saban
show