Scholar
Austin Xu
Google Scholar ID: OUw3iQgAAAAJ
Salesforce AI Research
Citations & Impact
Citations
203
H-index
9
i10-index
9
Publications
20
Co-authors
2
list available
Publications
Resume (English only)
Academic Achievements
- Published multiple papers in arXiv, EMNLP, ICML, and ACL, including 'Foundational Automatic Evaluators: Scaling Multi-Task Generative Evaluator Training for Reasoning-Centric Domains' and 'Hard2Verify: A Step-Level Verification Benchmark for Open-Ended Frontier Math'.
Research Experience
- Research Scientist at Salesforce AI Research. Interned at Duolingo and Amazon during his graduate studies.
Education
- PhD from Georgia Tech, advised by Mark Davenport and collaborated closely with Ashwin Pananjady. BSE from the University of Michigan. During graduate school, interned at Duolingo, working with Will Monroe, and at Amazon, working with Arjun Seshadri, Mariya Vasileva, and Achal Dave.
Background
- Research Scientist currently at Salesforce AI Research. His current research broadly focuses on improving the reasoning ability of foundation models, particularly large language models, with a particular emphasis on automatic evaluation (generative reward models/LLM-as-judges). He works largely on post-training and evaluation. Broadly, he is interested in the role humans play in the era of large models: when are human responses necessary, and when can we avoid collecting human feedback?
Miscellany
- In his free time, he enjoys cooking (and eating), reading, running, and watching basketball (NBA and college).
