CV
Academic CV for Zerun Niu.
Contact Information
| Name | Zerun Niu |
| zerun.niu@sydney.edu.au |
Experience
-
2026 - present Sydney, Australia
Casual Academic Tutor
The University of Sydney
Tutorial delivery and student learning support for university coursework.
- Delivered weekly tutorials, class activities, small-group discussions, and Q&A on course content and university learning platforms.
- Supported students in developing academic communication, teamwork, and professionalism skills through guided exercises and feedback.
- Created an inclusive tutorial environment to help students transition into university study and participate confidently.
- Coordinated with unit staff and followed teaching guides to run tutorials consistently and effectively.
-
2024 - present Sydney, Australia
Research Assistant
DUAL Group, The University of Sydney
Research engineering and experimentation for distributed learning and efficient AI systems.
- Built reproducible training and evaluation pipelines with configs, seeds, logging, and automated reporting.
- Supported submission and revision cycles by implementing baselines, targeted ablations, and publication-quality figures and tables.
- Addressed reviewer feedback through additional experiments and clarity edits in method and experiment sections.
- Coordinated experiment ownership, timelines, and artifact handoff across code, plots, and tables to reduce duplicated effort.
- Designed and deployed the official DUAL Group website for members, publications, and research projects.
Summary
Education
-
2025 - present Sydney, Australia
Master of Philosophy
The University of Sydney
Computer Science
- Supervisor: Nguyen Tran
- Research focus: Federated Learning, Semantic Communication, and Efficient AI Systems
-
- 2025 Sydney, Australia
Bachelor of Advanced Computing
The University of Sydney
Data Science
- Graduated with Distinction.
- Final years WAM: 77/100
- Relevant coursework: Machine Learning, Deep Learning, Natural Language Processing, Distributed Systems
Publications
-
2025 Federated Deep Equilibrium Learning over Resource-Constrained Edge Networks
Projects
-
Adaptive Low-Rank Optimization for Federated Learning
Undergraduate thesis integrating DyLoRA into a federated learning pipeline.
- Reduced communication cost by 80% without loss of accuracy.
- Improved training stability across clients with reproducible PyTorch experiments.
- Result: 85/100, High Distinction.
-
DUAL Group Website
Designed and deployed the official website for the Distributed compUting, optimizAtion, and Learning group.
- Introduced group members, publications, and research projects.
- Website: https://dual-grp.github.io/website-dual/
Teaching
Skills
GenAI and Multimodal: LLM/VLM orchestration, agents, tool use, multimodal adapters, LoRA, alignment, safety guardrails, jailbreak defense, RAG, automated evaluation, and human-in-the-loop evaluation.
Efficient Tuning: LoRA, DyLoRA, PEFT workflows, quantization, distillation, and efficient adaptation under constrained compute budgets.
ML Engineering: PyTorch, Hugging Face Transformers, PEFT, scikit-learn, reproducible training pipelines, experiment tracking, modular pipeline design, and API development.
Distributed and Data Systems: Ray, Dask, Databricks, Spark, Hadoop, Airflow, FastAPI, latency and performance optimization, tracing, and observability.
Programming: Python, SQL, Java, C/C++, Git/GitHub, Linux, Jupyter, PostgreSQL, MySQL, MongoDB, Pandas, and NumPy.
Awards
-
2025 Adaptive Low-Rank Optimization for Federated Learning, score 85/100.