Guardrails for a publishing pipeline: reject bad input at the source, make silent failures loud, gate the deploy, and handle the lifecycle — unpublish, rename, schedule.

The unglamorous half of a publishing pipeline: cross-repo orchestration, idempotency, manifest-as-state, and the bugs that only surface once it runs unattended.

The implementation behind a publish-flag pipeline: cross-repo triggers, pulling a private vault, converting Obsidian Markdown, and the write-back loop to avoid.

Why I turned my Obsidian vault into a headless content source — write where you think, flag a note to publish, and let the site build itself.

A repeatable framework for technical decisions you won't regret — frame the trade-offs, decide for your stage, and write down the why.

How to make code review actually improve the codebase — automate style, focus on design, ask questions, and keep PRs small.

Lessons from leading backend teams — shifting from output to outcomes, sharing context, owning interfaces, and multiplying your impact.

What to instrument for observability — structured logs, RED/USE metrics, and distributed traces — and how to alert only on what's actionable.

How to design REST APIs that age well — versioning, pagination, consistent errors, and backward compatibility that won't break clients.

A backend engineer's mental model for database indexing — what to index, the leftmost-prefix rule, and why a query still does a full scan.