Progress you can measure
Every result is a delta against a baseline. You will know whether your work moved the number.
Careers
Build the lab that runs itself
We are a small team in Germany working on a hard, concrete problem: making the experimental loop autonomous, and proving it with numbers rather than narrative. If you like owning whole systems, this is the place.
Why join
Small team, whole stack
Full-stack problems
Compute, scheduling, models, evaluation and product live close together — the loop crosses all of them.
Time to think
We automate the repetitive part of research on purpose, so human hours go to the parts that need them.
Room to own things
Small team means real ownership: you take a subsystem from idea to something the loop actually uses.
Open roles
Apply
Apply
Apply
Apply
Four places to start
All roles are full-time and based in Germany. Apply by writing to careers@autoscientia.com — tell us what you have built and send whatever shows it best.
Research Engineer — Automated ML
Design and run the loop itself: search spaces, proposer prompts, candidate ranking, and the wiring that turns results into the next round of experiments.
Systems Engineer — Experiment Orchestration
Make thousands of concurrent experiments boring: scheduling, batching, checkpointing, preemption, cost ceilings and fault recovery at scale.
ML Engineer — Evaluation & Tooling
Build the harnesses that decide what counts as progress: automated scoring, baselines that hold still, reproducibility, and result reporting nobody has to interpret.
Research Scientist — Scientific Reasoning
Work on how systems form hypotheses, choose what to test next, and know when a result is worth trusting — the parts of the loop that are still hard.
Nothing matches? We would rather meet the right person than fill a slot. If you
have built something unusual that belongs in an autonomous lab, write to us anyway.
What we look for
Evidence over pedigree
Titles tell us little. We care about what you have made, how you reason about systems that fail in interesting ways, and whether you can tell the difference between a result and a story about a result.
- Something you built end to end that you can explain clearly — a system, a pipeline, a benchmark, a paper, a product.
- Comfort with measurement: you reach for a baseline before an opinion.
- Willingness to work across layers instead of defending a boundary.
- German or English working proficiency; we work in both.
Hiring process
01
02
03
04
Intro conversation
Thirty minutes about your work and what you want to own next.
Technical deep dive
We go through a system you built, down to the decisions and the things that broke.
Scoped exercise
A small, paid, time-boxed problem close to the real work — no unpaid take-homes.
Offer
A conversation about scope, expectations and terms. Questions welcome in both directions.
Help us make experimentation relentless
Write to careers@autoscientia.com with what you have built — links beat adjectives.