Roone Ruanaidh

Other Titles Held: Innovation Architect, IT Service Architect, Customer Success Manager

• Own the enterprise AI sandbox on AWS & GCP, enabling safe use of State data with AI models.
• Lead an AI Community of Practice across 13 agencies under a Governor’s Executive Order.
• Trained over 300 State employees on approved AI tools across macOS and Windows endpoints.
• Built a coalition that secured $1.4M in budget commitments to implement a unified ITSM model.
• Unified 14 cabinet strategic plans and 18 IT portfolios into a data model for legislative reporting.
• Led architects from 10+ domains in sequencing the enterprise zero trust implementation plan.
• Awarded by the Governor for Individual Excellence in Performance; promoted 3 times in 4 years.

Other Titles Held: Platform Architect

• Increased new hire volume by 18% and permanent-hire rate by 31% post-digital transformation.
• Recovered $150K in overdue payments and reconciled over $350K in accounts receivable.

COBOL Reverse Engineering(for RFP and AI-rewrite evaluation)

• Mapped a legacy system containing critical workflows and data with no source-code access.
• Inventoried 692 data tables and 44,823 data columns across 31 interfaces in under 10 hours.
• Delivered full architectural report and interactive HTML reports to CIO and IT system owner.
• Cost was ~$2.7K all-in, with only ~$95 of that being AI compute costs.

Financial Audit Reporting Agent

• Distilled a manual workflow across 2 SaaS platforms and an internal solution into an AI agent.
• Created a skill that equipped the agent with a CSV rule set and deterministic Python test.
• Validated the agent’s accuracy across 3 multimodal intake formats and 75 historical reports.
• Cost was ~$2K all-in and saved ~156 hours annually, completed in under 6 hours using AI.

Public Comment Intelligence for Agency Commissioners

• Built a Python ETL pipeline that structured 599 image-rendered PDFs into normalized JSON.
• Enriched JSON entries across 13 amendments with 7 customer-provided sentiment categories.
• Delivered an interactive HTML heat-map that filtered comments across 113 unique locations.
• Built an AI agent that queried the enriched JSON file directly to accompany the heat-map.
• Cost was ~$1.2K all-in and saved at least 20 hours per amendment cycle.

Envelope

Researching how predictable the cost of agentic work becomes when a model can invoke programmatic capabilities that advance a task outside its context window.

University of Wisconsin–Madison

• Bachelor of Arts — Religious Studies, Ritual and Meaning Structures
• Bachelor of Arts — English, Creative Writing, Poetic Forms and Syntax