Tong Zhao |

Senior Manager, Post-Training Research at Scale AI Contacts: Personal E-mail: tong \(\alpha\tau\) tzhao \(\delta o\tau\) io Work E-mail: tong.zhao \(\alpha\tau\) scale \(\delta o\tau\) com School E-mail (deprecated): tzhao2 \(\alpha\tau\) nd \(\delta o\tau\) edu Profiles: GitHub, LinkedIn, Google Scholar.

I lead the post-training research team at Scale AI, where we build post-training data, methods, and agentic RL environments; develop frontier evaluations and benchmarks; and run post-training data and evaluation programs with leading AI labs. My research centers on scalable supervision for long-horizon agents in professional, scientific, and personal domains, as well as recursive self-improvement and agent safety and alignment. Previously, I was a Staff Research Scientist at Snap Research, where I led research and production efforts on agentic search, generative recommendation, and personalization. I received my Ph.D. in Computer Science and Engineering from the University of Notre Dame in 2022, advised by Prof. Meng Jiang.

News

• Oct. 2026: We released Humanity's Sixth Sense (HSS), a benchmark for intuitive visual reasoning in multimodal models!

• Sept. 2026: Our RSI-Bench is calling for task contributions from ML researchers!

• June 2026: I've joined Scale AI to work on post-training.

Honors and Awards

• Best Paper Award, ACM CIKM. 2025.

• Amazon Post-internship Fellowship, Amazon. 2021.

• Snap Research Fellowship, Snap Inc. 2020.

• Best Paper Award, DLG-KDD. 2020.

• SIGIR Student Travel Grant, ACM CIKM. 2020.

• Outstanding Teaching Assistant Honorable Mention, University of Notre Dame. 2019.