Dear Hiring Manager,
I build the tooling that removes the humans from the toil — AI agents that act, browser runners that never sleep, and MCP skill layers that give language models real reach.
I am writing to express my interest in joining [Company Name] as an automation and AI tooling engineer. My work sits at the intersection of AI agent design, developer tooling, and web automation: I author Model Context Protocol (MCP) servers and packaged skill libraries that extend Claude and other language models with first-class tool access — file systems, shell execution, web search, structured memory, IoT telemetry, and external APIs — all orchestrated through typed, composable interfaces. These are not wrappers; they are production-grade tool chains with error contracts, rate-limit handling, and predictable schema outputs that a language model can reliably call inside a multi-step reasoning loop.
On the browser automation side I use Playwright and Puppeteer for end-to-end extraction, interaction testing, and headless workflow execution. I have built pipelines that navigate authenticated UIs, scrape structured data across paginated or JavaScript-heavy surfaces, and emit clean output into downstream APIs or databases — with Chromium fingerprint controls and session persistence where stealth matters. I wire these runners into GitHub Actions or Docker Compose so they run on schedule, on event, or on-demand inside a CI gate. When the task needs a real-browser visual assertion or a pixel-diff regression test, I write that too.
What I find compelling about this domain is that good automation requires the same rigour as good software: clear contracts, observable failure modes, and the discipline to stop automating at the boundary where a human decision is genuinely needed. I apply that standard whether I am writing a Playwright scraper, designing a multi-step MCP tool chain, or wiring an agent skill that invokes shell commands on behalf of a language model. Security is not a finishing step — input sanitisation, permission scoping, and credential isolation are part of every tool I ship.
I work remotely from Indonesia and have done so since 2018, with a written-first, async collaboration style and a public record of the work at 1999azzar.github.io/1999AZZAR. My rate is USD 35/hour with a USD 250 project minimum, negotiable on scope, duration, and retainer shape. If you are building an agent platform, a developer tooling layer, or a browser-automation pipeline and want someone who has already shipped those pieces — I would like to hear about your stack and the next thing you are trying to automate.