cv

TBD

Basics

Name	Xianjun Yang
Label	AI Researcher
Email	xianjun.yang.ai@gmail.com
Url	https://xianjun-yang.github.io
Summary	AI researcher working on Agentic RL, mid-training, AI safety, and LLM alignment. Member of Technical Staff at Reflection AI. Completed PhD in CS at UC Santa Barbara in 3 years without prior CS background. ~3k citations.

Work

2026.02 - Present
Member of Technical Staff

Reflection AI

Working on open intelligence for large language models.
- Agentic RL
- Mid-training
2024.12 - 2025.12
Research Scientist

Meta Superintelligence Lab

Research across FAIR, GenAI teams.
- LLM alignment
- Mechanistic interpretability
2024.06 - 2024.10
Research Intern

Meta GenAI

Research internship at Meta GenAI, Seattle.
2024.03 - 2024.06
Research Intern

Amazon

Research internship at Amazon, Palo Alto.
2023.01 - 2023.06
Research Intern

NEC Laboratories America

Research internship at NEC Labs America, New Jersey.
2022.06 - 2022.09
Research Intern

Tencent AI Lab

Research internship at Tencent AI Lab, Seattle.

Education

2021.10 - 2024.11
PhD

University of California, Santa Barbara

Computer Science
2020.01 - 2021.09
PhD (Quit)

University of California, Santa Barbara

Fluid Mechanics
2015.09 - 2019.06
B.S.

Tsinghua University

Energy and Power Engineering

Awards

2026.01.01

ICLR 2026 Oral

ICLR

Oral presentation for 'Verifying Chain-of-Thought Reasoning via Its Computational Graph'.
2024.07.01

ICML 2024 Oral

ICML

Oral presentation for 'A Safe Harbor for AI Evaluation and Red Teaming'.

Skills

	Research Areas
	Agentic RL
	Mid-training
	AI Safety
	LLM Alignment
	Mechanistic Interpretability
	LLM Security

Languages

	Chinese
	Native speaker

	English
	Fluent