Projects
Selected Work
Serel Memory: Portable Memory for AI Coding Agents
AI coding agents forget everything between sessions. Serel Memory keeps project context in version-controlled markdown, so an agent picks up where it left off — portable across tools, reviewable in a pull request, and owned by you instead of a vendor.
MCP Integration & AI Governance Program
Deployed Model Context Protocol integrations across the internal tool stack at a consumer-lending company — Jira, Confluence, and more — then built custom MCP servers for the systems no off-the-shelf connector supported. Paired with an AI vendor evaluation framework defining assessment criteria, classification tiers, and approval chains, so adoption could scale without accumulating governance debt.
Internal program — case study, no public code.
Stakeholder Reports Hub
Built with Claude Code and deployed as a static site on AWS S3. Combines per-sprint receipts of completed, rolled-over, and dropped work, a weekly project newsletter, and a bi-weekly review of upcoming work — replacing status meetings with self-serve visibility for leadership and cross-functional partners.
Internal program — case study, no public code.
Release Notes Automation
GitHub Actions plus the Claude API, generating release documentation automatically on a daily release cadence. Manual release notes, eliminated.
Internal program — case study, no public code.
Foundations
This Site
Developed a modern, responsive personal website using Next.js and React. Features include a dynamic blog system with MDX support, dark mode toggle, and responsive design.
Cloud Resume Challenge
The full Cloud Resume Challenge, front to back: a serverless site on S3, CloudFront, and Route 53 with a DynamoDB/Lambda/API Gateway visitor counter, then an AWS CDK and GitHub Actions pipeline to deploy it. How I got hands-on with AWS beyond the exam.