Cloud Infrastructure as a Solo Engineer

The problem

Running a production SaaS alone, with real users, without a DevOps team or a senior engineer to lean on. Every infrastructure decision had to be correct the first time or cheap to reverse, because there was no one else to fix it.

The approach

I owned the full infrastructure lifecycle across GCP and AWS: provisioning, deployment, monitoring, secrets management, security, and handoff procedures. The goal was boring reliability — infrastructure that stayed out of the way so product work could continue.

What was built

• Multi-cloud provisioning on GCP and AWS
• Docker-based service isolation and consistent deployment environments
• Automated CI/CD through GitHub Actions with environment separation
• Secrets management through GCP Secret Manager — no credentials in code, rotation procedures defined
• Stress testing to validate 1,000 concurrent user capacity
• Security audit procedures and clean credential handoff processes
• Sentry-based error tracking and alerting across services

Migration work

Managed a Firebase migration alongside the frontend engineer, moving from managed services to more controllable infrastructure where it made sense for cost and reliability.