AI Research Intern, Seraph, Canggu, Bali

AI Research Intern — EX Venture — Canggu, Bali

AI Research Intern, Seraph, Canggu, Bali

Seraph research cluster · Seraph · On-site, Canggu, Bali · 4–6 months

What does it take to make multiple AI agents reason together on-device, without a cloud to hide behind, and still arrive at an answer you can check? This is an AI Research Intern building on-device multi-agent reasoning for the Seraph research cluster, a funded, multi-partner AI research cluster with UTN Nürnberg among the partners, at Seraph, a sciences company building real-world systems from one of nature's most versatile plants, lemna. The work sits inside a 256-paper knowledge base across 31 research areas, a working species-vision detection system that reads a whole cultivation tray, and first pilot trials at TRL 3. We are early, and we say so. What you leave with is the proof employers actually hire on: a working research artefact you built under real constraints, not a class project. Everyone has a degree. Almost nobody has proof.

What you build

Not a task list. The thing you build.

You build the on-device reasoning harness that lets multiple agents coordinate without a cloud round-trip. You work on the evaluation set that exposes where the agents disagree, hallucinate, or stall. You are on the hook for the reproducibility file, the one another researcher can run without you in the room. You build the benchmark report that names what failed and why, with the numbers attached.

The bar

A model or agent system you built outside coursework, with the failure mode under real constraints named. A training or evaluation pipeline another person can run without you in the room. Kept a hard reasoning problem on the table until the output was checkable. You can read a paper and tell which result would not survive contact with a different dataset.

Who this is for

Researchers and engineers who treat AI tools as instruments, not answers. You drive Claude Code, Hermes, or Codex every day and catch what the AI gets wrong before the result ships. The map is dead: the old path of waiting for a lab to anoint you is not where the interesting work is. A failed experiment is data, not failure. The work continues until the result is reproducible.

Where you do it, and why here

Thirty years building companies, New York, London, Tokyo, the Valley, and then it stopped mattering where the work happened. So the program runs on-site in Canggu, Bali. A modern research cluster costs twelve people and a toolchain, so being wrong six times a year is part of how we build. The campus is 4,000 sqm, and the people who run the companies are on it every day. Epic Solutions PT runs the on-site program in Canggu, Bali. What this location buys is attention, not a view.

What happens after

Seventeen interns have moved into full-time roles across the group. The internship is step one; paid work is the destination, for the ones who stay. The destination role and its terms are agreed before you commit, never after. We are early, and we say so, the path is written, not the seat.

Terms

Unpaid under Indonesian law. A room is available on campus, rent paid by you. Epic Solutions PT hosts the on-site program and assists with the C22B internship visa process. Daily lunch on us. Starting October. Questions about the fit or the timeline? Tell us, we'll talk you through how the program actually runs.

Requirements

A degree in computer science, engineering, or equivalent, you have proven you can finish something. You drive Claude Code, Hermes, or Codex every day and catch what the AI gets wrong. Something you built that nobody asked you to build: a model, an agent, a benchmark, a tool. On-site in Canggu, the full phase.

How to apply

Apply via the LinkedIn application field. Submit your CV and the repository.

A finished CV was never the point. What counts is a thing you've built and the drive to build the next one. That is who we are looking for.