Robin Jia

Robin Jia is an Assistant Professor in the Thomas Lord Department of Computer Science at the University of Southern California (USC), where he leads the AI, Language, Learning, Generalization, and Robustness (Allegro) Lab. His research focuses on enhancing the reliability and robustness of large language models (LLMs) through mechanistic understanding, benchmarking under distribution shifts, and neurosymbolic integration. Key affiliations include collaborations with the USC Keck School of Medicine and contributions to legal frameworks like the EU's Digital Services Act.

Robin's research spans multiple domains, including:

• Scientific analysis of LLM capabilities in in-context learning, data memorization, and numerical reasoning
• Advancements in robust NLP systems, emphasizing uncertainty estimation and calibration
• Development of methods combining LLMs with symbolic solvers for complex reasoning tasks
• Interdisciplinary applications in medicine and law, such as privacy-preserving synthetic data generation and medical misconception evaluation.

His recent publications (2024-2025) address Fourier-based numerical embeddings (NeurIPS), neurosymbolic planning (NAACL), and multimodal benchmarking (COLM), with a strong emphasis on privacy, fairness, and transparency. Scientific awards include the Google Research Scholar Award (2023), SoCalNLP Symposium Best Paper Awards, and ACL/EMNLP outstanding papers. He advises PhD students like Johnny Wei and Ameya Godbole, and has secured grants from the NSF, USC-Capital One, and USC-Amazon.