Experience

• Stood up the company's AI function on my own: ran the AI planning meetings, owned the infrastructure roadmap, and was the point person for all 19 GitHub Copilot licenses, interviewing every holder one on one and running a hub-and-spoke champion network across the dev team, Support, QA, Legal, and product. Four senior developers each paired with me on a workstream.
• Built two production agents that encode unwritten team knowledge: one for SQL conventions that generated scripts and queries already matching them (the package that used to mean copy-pasting from three or four old scripts), and one for WCAG compliance auditing that scanned the codebase, generated human-reviewed patch files via the SVN MCP server, and logged each change for audit. Validated against WAVE on the first of 50+ systems, cutting a 50-70 hour manual audit to about 10.
• Proposed and built the company's first MCP servers, three in total: a client-import automation tool (in development, targeting a 40-50 hour manual process), a database schema Q&A tool, and an SVN tool for codebase history and commit-quality Q&A.
• Led an initiative with AWS engineers to scale that MCP infrastructure via AgentCore, Bedrock, and Lambda into the reference architecture for that client-facing MCP server, including a proposed PDF-to-CSV pipeline pending AWS access.
• Interviewed nontechnical teams (PM, QA, Support, executive leadership) to find workflow bottlenecks AI could fix, then shipped three: a QA worksheet tracking pass/fail results, a support-ticket triage workflow, and a personal Teams agent that triaged my own inbox each morning.
• Contributed the AI angle to the company's UI/UX initiative: built CSS-focused AI skills that generated reusable classes and tokenized colors, rebuilt the internal site's frontend in three days, and led the formal design system with two junior developers. A working prototype is heading to a client reveal, drawn from an EMR product UI that hadn't been redesigned in about 20 years.
• Used AI to surface KPIs from exported data, a technique I demoed for our CEO that got me tasked with building a client-facing version, from spec to prototype.
• Proposed and mentored a 5-developer senior capstone team on a medication-adherence app prototype with AI chatbot functionality, drawing on my subject-matter expertise on the live client-facing product; the team placed 2nd; one was hired into a junior role I still mentor a year later, and I gave references to two others.

Writeups: AI adoption across teams · SQL convention agent · ADA compliance agent · MCP dev servers · First MCP server · Scaling on AWS · Design system · Capstone mentorship