News

Aug 20, 2026 LiveMacroEval is accepted to EMNLP 2026 Main! We release a live, contamination-resistant benchmark for LLM agents nowcasting 16 U.S. Macro indicators.
Aug 12, 2026 🥳 Honored to be selected as a recipient of the Workday AI PhD Fellowship! Many thanks to the Workday AI Research team for supporting my work.
May 16, 2026 IPO-Mine is accepted to KDD 2026 D&B Track! We release a toolkit and large-scale multimodal dataset for section-structured analysis of long, multimodal IPO filings.
May 11, 2026 🎉 I will be joining the Amazon Rufus Post-training team as an Applied Scientist Intern this summer. See you in Palo Alto!
Sep 18, 2025 WCB lands at NeurIPS 2025 D&B Track! We release the largest monetary policy corpus to date to benchmark how LLMs read what the world’s central banks are really saying.
Jul 07, 2025 Our COLM 2025 paper is out! We expose LLM’s financial knowledge amnesia: forgetting historical financials while hallucinating most on companies they sound surest about.
May 16, 2025 SiDyP is accepted to KDD 2025 Research Track! We use simplex diffusion with dynamic priors to denoise LLM-generated labels, lifting BERT classifier by 7%+ on auto-labeled NLP tasks.