Jason Xurong Liang is a final-year Ph.D. candidate in Computer Science at the University of Queensland, Australia. His research focuses on efficient recommender systems, lightweight embeddings, graph collaborative filtering, and LLM-based recommendation.
He is advised by Assoc. Prof. Rocky Tong Chen and Prof. Hongzhi Yin. He received his Bachelor of Computer Science with First-Class Honors from the University of Queensland in 2022.
Jasonβs broader research interests include data mining, recommender systems, user modeling, natural language processing, large language models, and scalable machine learning. His work has been published in venues including SIGIR, RecSys, TOIS, ICDM, and KDD.
From Oct 2024 to May 2025, Jason worked as an Applied Scientist Intern at Amazon, where he conducted research on LLM-based in-context learning for user cold-start recommendation in a commercial video streaming platform. This work was published at RecSys 2025. During the internship, he also built AWS-based ML/data pipelines using services such as EMR, SageMaker Studio, Bedrock, and S3 for large-scale user interaction data processing, distributed analysis, and downstream ML modeling. He worked with Dr. Julien Monteil, Dr. Vu Nguyen, Dr. Vuong Le, and Dr. Paul Albert.
His thesis submission is expected in September 2026, followed by oral defense in late 2026. He will be available for full-time opportunities from January 2027.
When he is not immersed in research π or in front of a computer π¨βπ», he enjoys traveling ποΈ, fishing π£, and watching movies πΏ.
π Publications
Under ReviewEmpowering Compact LLMs with Fusion of Layer-wise Exits for Recommendation, Xurong Liang, Tong Chen, Quoc Viet Hung Nguyen, Jianxin Li, Xiangliang Zhang, Hongzhi YinRecSys 2025In-context Learning for Addressing User Cold-start in Sequential Movie Recommenders, Xurong Liang, Vu Nguyen, Vuong Le, Paul Albert, Julien MonteilTOIS 2025Lightweight Embeddings with Graph Rewiring for Collaborative Filtering, Xurong Liang, Tong Chen, Wei Yuan, Hongzhi YinSIGIR 2024Lightweight Embeddings for Graph Collaborative Filtering, Xurong Liang, Tong Chen, Lizhen Cui, Yang Wang, Meng Wang, Hongzhi YinKDD 2024Hate Speech Detection with Generalizable Target-aware Fairness, Tong Chen, Danny Wang, Xurong Liang, Marten Risius, Gianluca Demartini, Hongzhi YinICDM 2023Learning Compact Compositional Embeddings via Regularized Pruning for Recommendation, Xurong Liang, Tong Chen, Quoc Viet Hung Nguyen, Jianxin Li, Hongzhi Yin
πΌ Research & Industry Experience
- Jan 2023 - Present, Ph.D. Researcher, University of Queensland, Australia.
- Designed and optimized lightweight recommender-system architectures using graph neural networks, large language models, and mixture-of-experts techniques.
- Developed efficient algorithmic frameworks to reduce embedding storage footprint, memory consumption, and computational overhead for large-scale recommendation.
- Built model training and data-processing pipelines on high-performance computing clusters using Python, PyTorch, and Slurm.
- Published first-author research in top-tier venues including SIGIR, RecSys, TOIS, and ICDM.
- Oct 2024 - May 2025, Applied Scientist Intern, Amazon, Australia.
- Developed an LLM-based in-context learning framework for user cold-start recommendation in a commercial video streaming platform; the resulting work was published at RecSys 2025.
- Built an AWS-based ML/data pipeline using EMR, SageMaker Studio, Bedrock, and S3 to process large-scale user interaction data, perform distributed preprocessing and analysis, and support downstream ML modeling.
- Worked with Dr. Julien Monteil, Dr. Vu Nguyen, Dr. Vuong Le, and Dr. Paul Albert.
- Sep 2021 - Sep 2024, Casual Research Assistant, University of Queensland, Australia.
- Conducted experiments, data preprocessing, baseline testing, and algorithm implementation for explainable recommender systems and hate speech detection.
- Contributed to a hate speech detection framework published at KDD 2024.
- Jun 2021 - Jul 2021, Winter Research Scholar, University of Queensland, Australia.
- Conducted the project Dynamic Topic Modelling on Social Media, using natural language processing to model topic changes on social media during the first 100 days of COVID-19.
- Supervised by Prof. Zi (Helen) Huang and Assoc. Prof. Rocky Tong Chen.
- Project visualization: https://topicvis.nifu.me/
π Education
- 2023 - Late 2026, Doctor of Philosophy in Computer Science, University of Queensland, Brisbane, Australia.
- Ph.D. project: Toward Efficient Lightweight Recommender Systems
- Thesis submission expected in September 2026; oral defense anticipated in late 2026.
- 2022, Bachelor of Computer Science with Honors, University of Queensland, Brisbane, Australia.
- Honors project: PEP++: Improved Pruning-Based Embedding Dimension Optimization for Recommender Systems
- Awarded First-Class Honors with cumulative GPA 7.000/7.000.
- 2019 - 2021, Bachelor of Computer Science, University of Queensland, Brisbane, Australia.
- Major in Data Science.
- Graduated with cumulative GPA 6.792/7.000.
- 2016 - 2018, High School Diploma, Indooroopilly State High School, Brisbane, Australia.
π§βπ« Teaching
- Feb 2022 - Jul 2026, Head Tutor & Teaching Assistant, INFS7901 β Database Principles, University of Queensland.
- Delivered practical and tutorial sessions, managed course administration, and coordinated assignment and examination marking.
- Helped students understand database design, relational databases, SQL, and related data-management concepts.
- Feb 2023 - Jul 2026, Head Tutor & Teaching Assistant, INFS3208/INFS7208 β Cloud Computing, University of Queensland.
- Delivered practical and tutorial sessions on cloud computing, distributed systems, and data-processing pipelines.
- Mentored student projects using Docker, Docker Swarm, Kubernetes, Spark, HDFS, Amazon S3, Google Cloud Storage, MongoDB, Redis, and Faiss.
π Awards
- 2022 Deanβs Commendation for Academic Excellence β Bachelor of Computer Science (Honors), University of Queensland, Australia.
- Awarded for achieving a grade point average of 6.60/7.00 or higher in a single semester.
- 2019 - Jul 2021 Deanβs Commendation for Academic Excellence β Bachelor of Computer Science, University of Queensland, Australia.
- Awarded for achieving a grade point average of 6.60/7.00 or higher in a single semester.
- 2021 Cloud Computing INFS3208 Best Project Runner-up Award, University of Queensland, Australia.
- 2020 Prentice Scholar, University of Queensland, Australia.
- Awarded to undergraduate Computer Science and Information Technology students whose academic performance ranked within the top 5% of their cohort.
- 2017 Subject Prize Award β Information Processing and Technology, Indooroopilly State High School, Australia.
- Awarded for ranking 1st in the course cohort.
π Scholarships
- 2023 - 2026 Australian Government Research Training Program (RTP) Scholarship
- 2024 Google Research Travel Scholarship
- 2022 UQ International Onshore Merit Scholarship
- 2021 Winter Research Scholarship
π§° Skills
- Programming: Python, SQL, Java, C, Haskell, Dafny, Scala, MATLAB, Shell/Bash, LaTeX
- Machine Learning & AI: PyTorch, Faiss, Scikit-learn, HuggingFace, PyTorch Lightning, NLTK
- Big Data & Databases: Spark, HDFS, MongoDB, Redis, relational databases
- Cloud & Infrastructure: Docker, Kubernetes, AWS, GCP, OCI, S3, GCS, Slurm/HPC
More about meβ¦
Want to know more? Please have a look at my CV π.
Webpage template cloned from acad-homepage.