AI/ML Intern

THE SITUATION

The Enterprise Architecture Office needed better ways to review architecture inputs, ground AI support in real portfolio context, and understand the software application portfolio across ministries and clusters. I worked across AI configuration, dashboard development, and product execution so the work became something people could actually use, not just a prototype.

WHAT I DID

• ▸Configured Azure AI Foundry models to support drafting, summarization, and structured architecture review inputs inside the workflow.
• ▸Used RAG techniques to ground AI responses in relevant architecture and portfolio context instead of generic model output.
• ▸Built Power BI and Microsoft Fabric dashboards for the EAO Director and Application Portfolio team to analyze the software application portfolio across ministries and clusters.
• ▸Owned product scope, roadmap, Jira backlog, milestones, and release planning while staying hands-on with the AI and dashboard work.
• ▸Standardized review artifacts and automated manual handoff steps across 3+ organizational branches.

IMPACT

• ✓Reduced architecture review cycle time by 30%
• ✓Removed manual handoffs across 3+ branches
• ✓Gave EAO leadership and the Application Portfolio team clearer visibility into the software application portfolio across ministries and clusters
• ✓Moved AI support from idea to working Azure AI Foundry configuration inside the review workflow
• ✓Made AI outputs more useful by grounding them with RAG-style context

WHAT I LEARNED

• ◆AI product work gets stronger when model configuration, retrieval context, dashboards, and workflow design are built together.
• ◆RAG is only useful when the source context and user journey are clear.
• ◆Dashboards work best when they answer real leadership questions, not just display data.
• ◆Product ownership helped me ship the technical work instead of leaving it as a prototype.

CHALLENGES I FACED

• ⚡Balanced hands-on AI configuration with product ownership responsibilities across roadmap, backlog, and stakeholder alignment.
• ⚡Worked with portfolio data that needed to make sense to technical architects, the EAO Director, and application portfolio stakeholders.
• ⚡Kept AI outputs useful, grounded, and appropriate for a public-sector architecture workflow.

FUN FACT

"The best part was seeing the same work support both AI-assisted review and leadership-level portfolio visibility."