About
I am Indra Prasetya, a senior software engineer. I build backend systems, the databases underneath them, and the pipelines that ship and check them. Most of my work is in Go, TypeScript, Python, and PostgreSQL.
This journal is where I write down what I learned while doing that work: migrations that had to run without downtime, services that failed in a way worth explaining, tooling for AI coding agents that I maintain, and the measurements behind any number I quote. If I cannot reproduce a figure, I do not publish it.
What I work on
- Backend engineering. Service design, APIs, idempotent job pipelines, and the unglamorous parts that decide whether a system is reliable.
- Database engineering. PostgreSQL schema design, migrations, indexing, and query performance.
- DevOps and infrastructure. Reproducible builds, CI/CD, containerised deployments, observability, and self-hosted infrastructure.
- AI engineering. The practical side of coding agents: context budgets, instruction files, and verification.
- Open source. I maintain or contribute to tooling I actually run, including agentdiet, a CLI that measures the context an AI coding agent loads, and cfgaudit, where I added the rule that detects hidden instructions separated by long runs of blank characters.
How this site is built
The publication is a static site built with Astro on the open-source AstroPaper theme, which is used under the MIT licence and customised here.
Every article is a Markdown file in a Git repository. Publishing means opening a pull request, and GitHub Actions builds, tests, and deploys the result to Cloudflare Pages. There is no database, no admin panel, and no tracking script. Search is a static index built at deploy time, so it works without a backend.
Elsewhere
- GitHub: github.com/prascli
- dev.to: dev.to/prascli
- Email: [email protected]
- RSS: /rss.xml
The writing here is published under CC BY 4.0. Quote it, translate it, argue with it, just link back.