I am a Data Science graduate from Deakin University, Melbourne, focused on data engineering and AI automations — building pipelines, orchestrating agents, and making data actually useful at scale.
My background spans data science and machine learning, but I find myself drawn equally to the infrastructure side. Right now I am deep in the Microsoft Fabric + Azure ecosystem, exploring how AI agents can plug into real data workflows — not as demos, but as systems that actually run. I also build mobile applications with Flutter and have a live app on Google Play.
I believe the most impactful work sits at the intersection of solid data engineering, intelligent automation, and clear product thinking. I learn in public — expect experiments, rough edges, and occasional pivots.
Open To
Languages
Frontend & Mobile
Data, ML & AI
Cloud, DevOps & Tooling
| Domain | Proficiency | Details |
|---|---|---|
| Multi-Agent Systems | ███████░░░ Advanced | Microsoft Agent Framework, Azure OpenAI, sequential/parallel/manager-driven orchestration |
| Data Engineering | ███████░░░ Advanced | PySpark, Delta Lake, Medallion Architecture, real-time streaming, Microsoft Fabric |
| Machine Learning | ███████░░░ Advanced | Classification, ensemble methods, class imbalance handling, cross-validation, feature engineering |
| LLM & RAG Pipelines | ██████░░░░ Proficient | Azure OpenAI integration, LLM tool use, retrieval-augmented generation, automated workflows |
| Data Visualisation | ████████░░ Expert | Power BI (DAX, DirectLake, MTD/MoM KPIs), R Shiny, Plotly, ggplot2 |
| Time-Series Forecasting | ██████░░░░ Proficient | Facebook Prophet, demand forecasting, real-time analytics with KQL |
| NLP | █████░░░░░ Proficient | Text classification, pipeline integration, Azure Cognitive Services |
| Mobile Development | ███████░░░ Advanced | Flutter, Firebase, Google Play deployment, AI-powered features |
⬡ Fitness Coaching AI Platform — Multi-Agent Coaching System
A next-generation coaching platform built on the Microsoft Agent Framework that separates domain expertise from execution flow. Six specialised agents collaborate through composable workflows to deliver safe, adaptive, personalised guidance — across onboarding, program design, progress tracking, and dynamic coordination. Safety validation is a hard gate before any plan reaches a member.
| Attribute | Detail |
|---|---|
| Stack | Azure OpenAI · Microsoft Agent Framework · Python |
| Architecture | 6 agent roles · 5 orchestration workflows · 3 coordination patterns (sequential, concurrent, manager-driven) |
| Scale | Designed for extensible multi-member concurrent coaching sessions |
| Safety | Dedicated safety_agent validates contraindications and load before every plan delivery |
| Impact | Decouples domain logic per agent, enabling independent team development and auditable safety compliance |
| Repository |   |
The platform implements three distinct orchestration patterns: sequential (strict agent ordering for onboarding), concurrent (parallel execution for independent data collection tasks), and manager-driven (a coordinator agent dynamically selects specialists based on member complexity). New agents and workflows are added independently under agents/ and workflows/ without touching core orchestration logic.
⬡ HyperScale Retail Command Center — Real-Time Fabric Analytics
An end-to-end Microsoft Fabric solution that transforms data from a passive record into an active business asset. The platform bridges the gap between demand signals and inventory action using the Medallion Architecture, real-time streaming, and a 30-day ML demand forecast — replacing a 24-hour reactive reporting cycle with sub-minute intelligence.
| Attribute | Detail |
|---|---|
| Stack | Microsoft Fabric · PySpark · Delta Lake · Facebook Prophet · Power BI DirectLake · KQL Database · Data Factory · Fabric Activator |
| Scale | $8.30M historical revenue · 5M units sold · multi-country dataset (UK, Netherlands, EIRE, Germany, France) |
| Performance | Alert latency: <1 minute vs. 24-hour traditional lag · 30-day daily revenue forecast |
| Automation | 100% automated pipeline from ingestion through alert delivery |
| Impact | Converted a once-daily batch process into a continuous intelligence layer, enabling same-day stockout prevention and intraday rebalancing |
| Repository | |
The architecture follows a Medallion (Bronze → Silver → Gold) layering strategy. Raw event data is ingested via Data Factory into Delta Lake, transformed with PySpark through cleansing and aggregation tiers, and surfaced in Power BI via DirectLake for zero-copy, real-time query performance. Fabric Activator triggers automated alerts on stockout risk rules evaluated against live KQL streams.
⬡ Bank Loan Performance Dashboard — $435M Portfolio Intelligence
A three-tier Power BI executive dashboard designed to transform raw loan transaction data into a guided narrative for stakeholders — moving from high-level portfolio health through market segmentation down to individual transaction auditing in a single coherent interface.
| Attribute | Detail |
|---|---|
| Stack | Power BI · DAX · Power Query |
| Scale | $435.8M+ loan portfolio monitored · three dashboard tiers (Summary, Overview, Details) |
| Performance | Advanced DAX with MTD, MoM variance, and dynamic financial ratio calculations |
| Security | Row-level analysis for risk exposure monitoring and bad loan containment |
| Impact | Surfaced 13.0% MoM funding growth alongside 86.2% good loan rate; identified $65.53M bad loan exposure requiring targeted intervention |
| Live Report |  |
The Summary tier surfaces portfolio-wide KPIs and the Good/Bad loan split. The Overview tier contextualises those numbers through borrower segmentation — employment length, loan purpose (Debt Consolidation dominates with 16,000+ applications), and term preference (75.85% choose 60-month terms). The Details tier enables granular transaction-level auditing for compliance and risk review workflows.
Master of Data Science · Deakin University · Melbourne, Australia · Graduated June 2026
Completed a master's degree in data science with applied focus on machine learning, statistical modelling, and data engineering. Built projects that translate coursework into real systems — from clinical ML pipelines to enterprise-scale analytics platforms on Microsoft Fabric.
- Applied machine learning across supervised, unsupervised, and ensemble methods with rigorous validation practice
- Hands-on data engineering with PySpark, Delta Lake, and cloud-native pipeline orchestration
- Data visualisation and storytelling with Power BI, R Shiny, Plotly, and ggplot2
- Mobile application development with Flutter resulting in a live Google Play deployment
| Recognition | Details |
|---|---|
| 🚀 Live Google Play App | C-Meds AI Medication Reminder shipped and live on Google Play Store |
| 🏆 GitHub Pull Shark | Earned the Pull Shark achievement for consistent open-source contribution activity |
| 📊 End-to-End Fabric Solution | Built a full Medallion Architecture pipeline on Microsoft Fabric from raw ingestion to Power BI |
| 🤖 Multi-Agent System | Designed and shipped a 6-agent coaching platform on the Microsoft Agent Framework |
| 📈 $435M Portfolio Dashboard | Delivered a three-tier executive Power BI dashboard monitoring a $435M+ loan portfolio |
| 🌱 Learning in Public | 29 public repositories spanning ML, data engineering, mobile, and AI automation |
Microsoft & Azure
Data Engineering
Mobile
current_focus:
learning:
- "Medallion Architecture patterns — Bronze/Silver/Gold layer design at scale"
- "Real-time streaming with Microsoft Fabric and KQL for sub-minute analytics"
- "dbt for data modelling, SQL performance tuning, and pipeline observability"
building:
- "Multi-agent AI workflows on Azure OpenAI + Microsoft Agent Framework"
- "End-to-end Fabric pipelines: PySpark ETL → Delta Lake → DirectLake Power BI"
- "Extending C-Meds with smarter AI prescription parsing and scheduling logic"
exploring:
- "LLM tool use and RAG pipelines for real data workflow automation"
- "Pipeline observability — monitoring, alerting, and lineage for production pipelines"
- "ONNX and lightweight model serving for mobile and edge inference"
open_to:
- "Data Engineering and ML Engineering roles in Melbourne or remote"
- "AI Automation projects on the Azure / Microsoft Fabric ecosystem"
- "Flutter development and mobile AI opportunities"
- "Open source collaboration in data engineering and MLOps tooling"
"I don't just want data to be stored — I want it to work."