
Josh GardnerResearcher
Anthropic
Member of Technical Staff
via Anthropic
Researcher · Academic
Assistant Professor of Computer Science at Stanford University
Member of Technical Staff at Anthropic
Schmidt exposed how models that score well on familiar benchmarks can fail under natural changes in data. His group builds more reliable evaluation and training sets, including LAION-5B, OpenCLIP and DataComp, and studies how dataset quality shapes multimodal and language models.
See something inaccurate or outdated?
Stanford University
AnthropicWork and education history is primarily focused on AI-relevant roles and may not be comprehensive.
Last editorial review: August 10, 2026