I design connected BI and AI systems, from governed data infrastructure to stakeholder-facing solutions, and the agentic engineering workflows used to build them.
My career began in client consulting and commercial and research analytics. Power BI, Microsoft Fabric and Power Platform led me into BI engineering. Today I work across those layers as an architect, connecting stakeholders' needs with technical choices to drive business outcomes, while keeping future maintainability in mind.
At Mastercard, I designed and continue to evolve a dynamic departmental analytics infrastructure. I also coordinate the department’s Copilot Studio environment, building stakeholder-facing agents connected to enterprise knowledge and Fabric Data Agents through workflows and APIs.
Alongside this, I build reusable, specialist-agent GitHub Copilot workspaces for Fabric and Power Platform, plus emerging agent-led patterns for repeatable app development. The objective throughout is useful solutions that can be built, operated and improved over time.
What I build · Public work · Core domains · Tools & platforms · Connect
Stakeholder-facing BI solutions
End-to-end, multi-platform (Fabric/Power Platform) solutions: from automating data processes to building solutions that allow stakeholders to read and write data, while leadership can monitor the business.
Stakeholder-facing AI solutions
Copilot Studio agents connected to Fabric Data Agents to make enterprise information (business knowledge and data) accessible through purposeful AI experiences.
Agentic engineering systems
Multi-agent development workspaces for Fabric and Power Platform, alongside experiments with repeatable agent-led application delivery. The stakeholders here are developers and solution teams.
| Fabric Agentic Workspace — One-Click Setup | Power Platform Workspace — One-Click Setup |
|---|---|
| A reusable GitHub Copilot development workspace for Microsoft Fabric and Power BI. A structured 47-agent organisation, curated skills, MCP/CLI integrations and governed development guidance. | An agentic development workspace for Power Platform, with a 13-agent primary organisation, Microsoft skills, authoring tools and controlled DEV → TEST → PROD delivery. |
| Explore Fabric workspace | Explore Power Platform workspace |
Both are public pre-releases under active development.
DAX & Data Modelling
Reusable DAX patterns and semantic modelling techniques, including filter context, time intelligence, virtual tables and interactive controls.
Power BI Visual Templates
Reusable analytical visuals built from native Power BI components, including IBCS-style variance layouts and Direct Lake variants.
Power BI HTML Templates
Custom KPI cards and waterfall visuals built with DAX, HTML and CSS for Power BI.
Power BI Reports
Selected analytical case studies, competition entries and personal Power BI projects.
Architecture & engineering: BI Solutions Architecture · AI Architecture · BI Engineering · ETL & Data Pipelines · Semantic Modelling · Data Governance
Risk & delivery: Data Security · Access Management · Data Automation
Analytics & experience: Data Visualization · Data Storytelling · Predictive Analytics
Applied AI: Agentic Development · AI Agents Development · Agent Skills
Data platforms, BI & engineering: Microsoft Fabric · Power BI · SQL · Apache Spark · PySpark · DAX · Power Query M · Lakehouse · Data Warehouse · Fabric SQL Database · Fabric Data Pipelines · Semantic Models
Power Platform & automation: Power Automate · Power Apps · Copilot Studio · SharePoint
Web & interface development: HTML · CSS
Agentic AI development: Fabric Data Agents · GitHub Copilot · OpenAI Codex · Claude Code · Custom Agents · Agent Skills · MCP · Prompt & Context Engineering · CLI
Development & delivery: VS Code · GitHub · Azure DevOps · Git
- Winner — Maven Coffee Challenge
- Winner — Maven Slopes Challenge
Both achievements reflect an earlier chapter of my work in analytics and data visualization.
The BI and data community has been an important source of learning throughout my career. I share reusable patterns, tools and practical walkthroughs when they solve a real problem or offer a useful approach others can adapt.
I prefer depth over publishing frequency. Mature solutions, actively evolving tools and experiments should be described for what they are — not presented as the same thing.
