Welcome to Zhuang Li’s Personal Website
Welcome!
Hello! I am Zhuang Li, a Lecturer at RMIT University and formerly a Research Fellow at Monash University. My research focuses on large language model (LLM) safety, evaluation, and structured natural language processing (NLP), with an emphasis on building trustworthy and inspectable AI systems. I am particularly interested in making LLMs safer, more reliable, and easier to evaluate in real-world settings.
About Me
I earned my PhD from Monash University, where I was mentored by Dr. Lizhen Qu, Prof. Gholamreza Haffari, and ACL Fellow Phil Cohen. I am now a Lecturer at RMIT University.
My recent work has appeared in leading venues such as ACL, EMNLP, NAACL, NeurIPS, and ICLR. I was a recipient of the EMNLP 2025 Outstanding Paper Award, one of seven such awards among 3,200+ accepted papers.
I have also collaborated with industry partners, including Ant International (Singapore), on LLM safety evaluation and red teaming, resulting in publications at ACL and NeurIPS.
I also contribute to the research community through service roles, including Senior Area Chair for EMNLP 2026, Area Chair for NeurIPS 2026, ACL Rolling Review, and NLPCC 2026, and Publication Chair for ALTA 2026.
I am dedicated to developing trustworthy NLP systems and welcome collaboration in LLM safety, evaluation, and structured approaches to language understanding.
News
04/2026: Three ACL 2026 papers, including one main conference paper and two Findings papers, on LLM steering, alignment, and robustness were accepted.
03/2026: I am a co-applicant on the Wellcome Trust AI4You(th) project on safe clinical LLMs for youth mental health, an approximately AUD 6M grant jointly secured by Orygen, the University of Melbourne, and RMIT University, in partnership with Google Health.
03/2026: Serving as Senior Area Chair for EMNLP 2026.
11/2025: Our paper DiscoSG: Towards Discourse-Level Text Scene Graph Parsing through Iterative Graph Refinement received the EMNLP 2025 Outstanding Paper Award, one of seven such awards among 3,200+ accepted papers. Built on our previous FACTUAL work and integrated into FactualSceneGraph, which now has 240,000+ PyPI downloads and 600,000+ model downloads on Hugging Face. Links: DiscoSG paper · FACTUAL paper · DiscoSG repo · FactualSceneGraph toolkit
10/2025: One paper on LLM safety was accepted to NeurIPS 2025, and our workshop proposal PersonaLLM was also accepted. Link: PersonaNLP workshop
09/2025: Two EMNLP 2025 papers were accepted, including one in the main conference and one in Findings.
05/2025: Six papers were accepted to the ACL 2025 main conference.
05/2025: Served as an Area Chair for ACL Rolling Review.
02/2025: Served as an Area Chair for ACL Rolling Review.
- 12/2024: I joined RMIT University as a Lecturer.
I will be recruiting one PhD student in the 2025 cycle to work on low-resource languages and culturally aligned LLMs.(filled)I may collaborate with researchers from RMIT University and Monash University to recruit another PhD student focusing on multimodal NLP.(filled)- If you are interested in joining, feel free to reach out.
- 04/2024: Co-organised the ALTA Shared Task: Overview of the 2024 ALTA Shared Task: Detect Automatic AI-Generated Sentences for Human-AI Hybrid Articles.
Research Interests
- LLM safety and evaluation
- Structured NLP and semantic parsing
- Efficient adaptation and data synthesis
Collaboration and Supervision
I am eager to collaborate on innovative NLP and machine learning projects, particularly in LLM safety, evaluation, red teaming, structured NLP, and trustworthy AI systems.
I welcome collaboration with:
- Remote interns with strong research backgrounds for collaborative projects
- Industry partners interested in practical challenges in LLM safety and trustworthy language technologies
For PhD supervision at RMIT, I collaborate with senior faculty members at RMIT as part of the supervision team. I welcome applications for:
- Grant-funded students (subject to grant availability)
- University scholarship applicants (highly competitive)
If this sounds relevant, feel free to send your CV and research interests to zhuang.li@rmit.edu.au.
Australian PhD Application Tips
Here is a practical overview for students who are considering doctoral studies in Australia and may be unfamiliar with the application process. These notes are based on my experience and are not official admission or scholarship rules.
Supervisor-Supported PhD Positions
Some PhD positions are linked to a supervisor’s funded project. If your background fits the project well, the supervisor may support your application as part of a grant-funded supervision team.
Useful things to prepare:
- A concise CV with education, research experience, publications, projects, and technical skills
- Transcripts and evidence of research training, such as an honours thesis, master’s thesis, research internship, or substantial project
- A short statement explaining why your interests fit the supervisor’s current research program
- English language evidence, such as IELTS, TOEFL, or PTE scores, if required by the university
Timing can be more flexible than central scholarship rounds, but funding availability, formal admission requirements, and English language requirements still apply. For RMIT, applicants should also check the official Higher Degree by Research (HDR) application and research scholarships pages.
University Scholarship Opportunities
For international applicants: University scholarship rounds are usually more competitive and deadline-driven. AI-related PhD applications can be especially competitive because many strong applicants are applying for a limited number of scholarships.
Competitive applicants often have several of the following strengths:
- Strong academic results, often around High Distinction level or equivalent in the applicant’s grading system
- Clear research experience, such as a thesis, research assistantship, internship, preprint, or peer-reviewed publication
- Evidence of independent writing and technical ability, especially for NLP, machine learning, or data-centric AI topics
- A focused research fit with the proposed supervision team rather than a generic interest in AI
- English proficiency evidence that meets the university’s current requirements
These are not fixed requirements. Scholarship outcomes depend on the full applicant pool, school ranking process, university rules, and available funding in that round. Always check the official university HDR scholarship pages for current deadlines and eligibility rules.
For Australian residents:
- Domestic applicants may be eligible for Research Training Program (RTP) places or scholarships
- Universities may also offer stipend scholarships or top-ups depending on the scheme
- Strong grades and evidence of research potential are still important
- Eligibility criteria, stipend amounts, and application timelines vary by university and year
Personal Interests
Outside academia, I enjoy hobbies that inspire creativity and relaxation, including gaming, movies, and reading.
My favourite books, movies, and games
Gaming: AAA games like Elden Ring, The Legend of Zelda: Breath of the Wild, The Legend of Zelda: Tears of the Kingdom, Diablo II, III, IV, Black Myth: Wukong, Red Dead Redemption, The Witcher 3: Wild Hunt, Cyberpunk 2077, God of War series, etc.
Movies: Sci-fi and fantasy classics like Star Wars, The Lord of the Rings, and Batman: The Dark Knight. Other popular favourites include the Indiana Jones series, The Shawshank Redemption, The Silence of the Lambs, Forrest Gump, the Harry Potter series, and The Chronicles of Narnia series.
Books: Works on history and economics, including Les Misérables and The Intelligent Investor.
I’m also active on Zhihu (@Alpaca).