Analytics & Intelligent Systems
- Machine Learning
- Quantitative Business Analysis
- Visual Analytics / Data Visualisation
- Managing Data at Scale
- Elements of Data Processing
I am Zimeng (Michelle) Liu, a University of Sydney student combining data analytics, finance and business analysis with research in RAG/GraphRAG, ESG intelligence and evidence-based decision systems.
Study profile: The University of Sydney ↗
My academic and professional work sits at the intersection of data, finance, strategy and responsible technology. I like problems where the evidence is messy, the stakeholders disagree, and the final recommendation still has to be measurable and defensible.
I identify most with the ENTJ profile: strategic, organised, decisive and comfortable taking ownership. I translate that into practical strengths rather than a label—setting direction early, structuring ambiguous problems, communicating under pressure and turning analysis into an executable plan.
I enjoy setting a clear hypothesis, breaking the problem into evidence streams, testing assumptions and communicating a concise recommendation. Debate, research and cross-disciplinary projects have strengthened my ability to challenge ideas without losing sight of execution.
Coursework is grouped by the capabilities it developed rather than by semester, reflecting the interdisciplinary way I approach problems.
Sep 2022 — Sep 2025 · Sydney, Australia
Led recruitment, training, events and inter-university competition participation. Represented the University of Sydney at the 2024 Planet Cup, where the team advanced to the final three. The role strengthened fast issue-framing, argument structure, public speaking and high-pressure Q&A.
My internship page expands each role into the problem, tools, analysis and outputs, with quantified results only where supported by the underlying work.
Open the internship timeline →Click a marker. The current version records locations; later each marker can open a story, project, exchange, competition or travel memory.
Projects span machine learning, industry research, finance, data systems, policy and ESG. Team projects are labelled as such.
View all projects →30 corporate reports, 2,015 benchmark questions and controlled comparisons across non-graph and graph-based retrieval systems.
Compared four telecom peers across six financial metrics and designed an 80/20 debt-reduction and special-dividend allocation for asset-sale proceeds.
XGBoost-led targeting of ~20,000 high-probability lapsed donors, with A$150K base-case projected revenue against A$45.4K estimated campaign cost.