Jamie Geddes

A bearer-authenticated API and a ever growing 26.7 TB S3 mirror that let an external research partner search a national archaeology archive and pull collections into their own virtual research environment.

• PHP
• Symfony
• API Platform
• Docker
• SQLite
• +3

An internal dashboard that turns the archive's reporting questions into one click, replacing SQL written by hand every time a stakeholder asked.

• Node.js
• Express
• React
• Vite
• Tailwind
• +2

Remote staff were dragging archive transfers back down the VPN across several network hops; moving the work inside the network removed the round trip entirely.

• Windows Server
• NFS
• VPN
• Networking

A live climbing shoe price comparison site — scraping five retailers, matching their listings to one product, tracking prices over time, and serving it through a faceted search on infrastructure I own end to end.

• React
• Vite
• Tailwind
• Express
• PostgreSQL
• +5

Every Oracle, PostgreSQL and MySQL backup at the archive was run by hand; now they run themselves, report to Slack, and have been restored from.

• Bash
• Python
• Oracle
• PostgreSQL
• MySQL
• +3

The company had a very out of date and limited disaster recovery plan. Writing one turned up the finding that mattered — database backups were being discarded after 90 days — and the retention is now twelve months.

• Cloudflare
• Linux
• NFS
• TLS
• MFA