About · Hannah Weng · Sydney
Hi, I’m Hannah
Cloud & AI Solution Architect — designing AI-enabled enterprise systems from workshop to production.
The left branch holds the years already lived — hover any leaf to read its story. The right branches grow one per year ahead, each leaf a post or milestone still to come.
that survives Monday morning.”
I'm a Cloud & AI Solution Architect based in Sydney, with seven-plus years across Microsoft Power Platform, AI engineering, Databricks, and Microsoft Fabric. I've worked with telecommunications, healthcare, education, and public-sector customers — running workshops, scoping PoCs, writing SOWs, and delivering production systems aligned to CAF and WAF principles.
My research roots run deep: a PhD in AI in time series prediction and deep learning, with peer-reviewed work in cellular traffic forecasting and biomedical video stabilisation. That research lens shapes how I architect — measure before optimising, keep a paper trail, and treat evals as a first-class artifact.
I write here mostly to think out loud. Architecture decisions, MLOps walkthroughs, Power Platform debugging, and the small lessons that took an afternoon to learn and deserve to be written down once.
Innovation Award
Designed and delivered a Microsoft Copilot & Power Platform AI agent that achieved a 70–80% reduction in retrieval time for a legal workflow — combining prompt engineering, Microsoft OpenAI, and Power Automate orchestration.
-
2023
Spatial Bottleneck Transformer for Cellular Traffic Prediction in the Urban City
Australasian Joint Conference on Artificial Intelligence (AJCAI), pp. 265–276. Springer Nature Singapore.
-
2023
Autonomous Stabilization of Retinal Videos for Streamlining Assessment of Spontaneous Venous Pulsations
45th Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC), pp. 1–4.
Agentic AI systems on Microsoft AI Foundry and end-to-end ML pipelines on Databricks — multi-modal classifiers, retrieval-augmented assistants, and the operational glue that turns them into reliable services.
Foundation models, agentic patterns, and the operational reality of running them — Microsoft Fabric, Snowflake, and the long tail of governance and compliance details that decide whether a PoC graduates to production.
Architecture diagrams, ADRs, post-mortems, and tutorials. Documenting what I build and what the field is shipping so future-me (and future stakeholders) don’t have to start from scratch.
Want to scope an AI engagement, review an architecture, or just compare notes?
Always open to interesting projects, second opinions on architecture decisions, or a chat about MLOps, Power Platform, or anywhere the two intersect.