Writing on backend, cloud and system design.
Simple, practical guides on building and running backend systems on AWS. Diagrams first, then code. For developers and the people who hire them.
- Zero-downtime Lambda deploys from GitHub Actions: OIDC, versions, aliases and canary rollback Deploy AWS Lambda from GitHub Actions with no stored AWS keys: OIDC login, versions, a weighted alias canary and automatic rollback.
- Self-hosting an LLM with vLLM: the real cost math vs an API A neutral break-even model for self-hosting an LLM with vLLM vs paying an API per token, with a worked example, real prices and a small production setup.
- Cache-aside, write-through and stampede protection: Redis in front of Aurora PostgreSQL Redis caching patterns for Postgres in plain words: cache-aside vs write-through, cache stampede prevention, TTL jitter and safe invalidation.
- Terraform for a one-AWS-account startup: layout, state, and the day you split How to structure Terraform for a small AWS startup: one state, then split state per stack and environment, then multi-account. With real HCL.
Follow along with the RSS feed.