Principal Cloud Data Architect
Data platforms
that survive
the audit.
Planning, architecture, and delivery for oil & gas — from source systems, through cloud warehouses, to AI that answers questions, writes summaries, and ships reports.
What I do
One architect,
the whole stack.
AI engineering
LLM and MCP-native tooling, SOX-compliant document-to-ERP pipelines — AI where it removes hours of manual work, not as a buzzword.
Warehousing
Snowflake estates: dimensional models, cross-reference systems, master data, dynamic tables, governance, and cost control.
Multi-cloud
Reference architectures and production estates across Azure, AWS, and GCP — security, budgeting, and administration included.
Oil & gas
Knows what a PROPNUM is. ERP, production, land, midstream, and drilling systems unified into reporting layers that tie out.
Planning & architecture
Reference architectures, costing estimates, deployment scripts, UAT documentation, and delivery leadership end to end.
Hot offerings
What teams are
hiring for right now.
AI integration
LLMs wired into the systems you already run — ERP, warehouse, and document workflows — with deterministic validation and audit trails, not demos.
AI strategy & consulting
Roadmaps, architecture reviews, and build-vs-buy decisions for teams adopting AI — grounded in platforms already running in production.
Chatbots & AI agents
Retrieval-grounded assistants and MCP-native agents over your own data — answering questions, writing summaries, and shipping reports.
Data extraction & pipelines
Documents, APIs, and legacy systems turned into clean, governed tables — the ingestion and engineering layer everything else stands on.
Analytics & reporting
Dimensional models and reporting layers that tie out to source — dashboards and scheduled reports your finance team can actually sign off.
AI-readiness audit
A fixed-scope review of your data estate and AI opportunities — findings, costed target architecture, and a delivery plan you can execute.
Training-data & evaluation pipelines
Labelling, review, and versioning workflows — so a model’s training data and its evaluations stay as governed and reproducible as the reports it feeds.
Featured work
Problems solved,
audits survived.
AI journal entry automation
Bank statements in, SOX-compliant ERP journal entries out. An AI pipeline with deterministic validation and full auditability, deployed across dual Azure tenants for a US oil and gas operator.
Read the case study →
MatchMaker AI
Five source systems, five naming conventions, zero shared keys. An ML matching engine built end to end in Snowflake and Python that turned disagreement into one cross-reference backbone.
Read the case study →
Wellbore schematics app
Drilling data turned into interactive 2D/3D pictures of what's actually downhole — generated from governed Snowflake data, so the schematic never drifts from the source system.
Read the case study →