About the role
Make AI-assisted product journeys more useful, reliable and testable.
What you’ll do
- Improve agent instructions, retrieval, tool use, evaluations and failure handling.
- Connect each change to a real user task and document its limitations. Start with scoped tasks alongside a mentor; production access, merges and releases remain with authorized maintainers.
What we need to see
- Python and AI workflow fundamentals.
- Structured debugging and the judgment to distinguish a convincing demo from a reliable system.
What winning looks like
- A focused agent improvement with a reproducible demo, test cases and a before/after comparison.
- Explain where it still fails and how you verified the result.
Who can apply
Students and recent graduates in relevant disciplines may apply. A specific degree title is not the main filter. Show useful work, clear thinking, the ability to learn from review and follow-through.
Where and how we work
We work in the office together five days a week, in one of our garages. Remote work is available only by exception, arranged individually around specific circumstances.
Engagement
Internship duration, manager, start date and performance matrix are stated in the individual written offer. A later full-time offer depends on sustained performance, readiness, graduation where applicable and business requirements.
Compensation
This is a paid internship. The stipend and the terms of any potential full-time offer are stated in your individual written offer.