Deskripsi pekerjaan AI Automation Engineer (Forward-Deployed) Surya Narayana Indonesia
About the role
We're an early-stage team that rebuilds businesses' most time-consuming workflows around AI, and keeps them running reliably in production. We work directly inside our clients' operations, finding the workflow that costs them the most coordination and rebuilding it to be faster, cheaper, and more dependable.
You'd be the person who builds those workflows. This is a hands-on engineering role with real ownership: you take a workflow we've selected with a client, work alongside the team to map how it actually runs, and build the automation that replaces the manual coordination — with proper logging, testing, review steps, and rollback built in from the start, not bolted on after.
It's a forward-deployable role, meaning you're close to the actual work and the actual client, not building in isolation. If you like seeing what you build get used in the real world and matter to a real business, this is that.
What you'd do
- Build AI-powered automation for real client workflows — agent orchestration, API integrations, connecting the systems a business already runs on.
- Work with the team to map how a workflow actually operates: the steps, the data it touches, the judgment involved, and a clear baseline to measure improvement against.
- Build reliability in from the start — testing, logging, human-review checkpoints, and safe rollback — so what you ship can be trusted in live operations.
- Prove each workflow in a shadow run alongside the existing process before it goes live, and refine it based on what the run surfaces.
- Document what you build to a standard where someone else could operate and maintain it .
- Monitor live workflows, catch problems before they surface as client complaints, and improve them over time.
What we're looking for
- Solid, self-directed engineering ability — you can own a build end to end without needing close supervision, and you make sound calls on your own.
- Genuine care for reliability and maintainability. You're the kind of engineer who documents as they go, builds things others can pick up, and doesn't cut corners that hide until later.
- Familiarity with automation and agent tooling — workflow platforms, agent/LLM frameworks, retrieval over documents, plus solid Python and Git — is welcome. We care more about how you reason about building something reliable than which specific tools you've used.
- Working understanding of how to make AI/LLM-based automation behave predictably: testing outputs, handling failure, knowing when a human needs to stay in the loop.
- Willingness to be close to the client and the real operation — comfortable being embedded in how a business actually works, not just heads-down in code.
Nice to have
- Experience taking automation into production and keeping it running
Why this role
You'd be an early member in a team at the start of its growth, with unusually direct ownership of the work and its impact. You'll be building things that go live and matter, not tickets in a backlog. As we grow, early builders grow with us.
