Akari Asai

I am an Assistant Professor at Carnegie Mellon University Language Technologies Institute, with an affiliate appointment in the Machine Learning Department and a research scientist at Allen Institute for AI.

Please read the FAQ before emailing me if you’re interested in working with me.

I build language models and agents that can go beyond their training data and operate reliably in the world, especially on open-ended, long-horizon tasks involving search, evidence synthesis, and interaction with external tools and environments.

• Beyond training data: Models that search and act. We develop models and agents that seek new information, use tools, and interact with external environments to solve complex, open-ended tasks.
  DR Tulu (ICML 2026 Oral) · Self-RAG (ICLR 2024 Oral) · AgentIR · Task-aware Retriever · DRAGON · PathRetriever

• Reliable AI in the wild. We study when AI systems fail as they interact with real-world information and tools, including hallucinations, copyright and licensing issues, and unreliable reasoning, and develop methods to improve reliability.
  Binary RAR (ICML 2026 Oral) · When Not to Trust LMs · FAVA · CopyBench · REALTIMEQA

• AI for Science. We build AI systems that help scientists search, synthesize, and reason over scientific knowledge, with the longer-term goal of supporting open-ended scientific research and discovery.
  OpenScholar (Nature 2026) · AI Reviewer · Into the Unknown · ScholarCatalyst

Across these directions, I am also interested in making language technologies work across languages and cultures, especially for global information access. XOR QA · CORA · AfriQA · PANGEA

Our work has been recognized by the AI2050 Early Career Fellowship, the AAAI/ACM SIGAI Dissertation Award, MIT Technology Review’s 35 Innovators Under 35, and Forbes 30 Under 30 Asia in Science, and featured in MIT Technology Review, Forbes, Nature News, and Science.

Previously, I completed my Ph.D. at the University of Washington in 2025, advised by Hannaneh Hajishirzi, where I worked on retrieval-augmented language models and information-seeking systems. My dissertation focused on building and advancing the foundations of retrieval-augmented language models (thesis, video), and received the AAAI/ACM SIGAI Doctoral Dissertation Award. Prior to UW, I received my B.E. in Electrical Engineering and Computer Science, where I received the Dean’s Medal.