Sovrano AI connects Europe's most qualified graduate talent to the post-training pipelines of the world's leading AI labs. As a Master's Fellow on the Expert Track, you are not a generalist annotator - you are a domain-credentialed evaluator whose judgment helps shape what frontier models know and how they reason.
Expert Fellows work directly on projects commissioned by the world's top LLM companies - the labs building the most capable and most widely deployed AI systems on the planet. You won't just learn about frontier AI from a textbook or a course. You'll be inside the machine, doing work that actually matters to the teams building it.
Expert track work is more demanding, not just better paid. You will be expected to produce structured reasoning traces, write detailed evaluation justifications, and meet quality benchmarks significantly higher than the generalist track. Do not apply if you want low-effort side income.
KEY RESPONSIBILITIES Evaluate LLM outputs in your domain of study - identifying errors, gaps, and reasoning failures that generalist annotators cannot catch Write step-by-step reasoning traces and ideal model responses Contribute to rubric design for your subject area Test model outputs for factual accuracy, logical consistency, and domain appropriateness Participate in structured feedback cycles with Sovrano AI's project leads IDEAL QUALIFICATIONS Master's students at a Sovrano AI partner university anywhere in Europe Strong academic record in a quantitative, scientific, or technical discipline Exceptional written English; ability to produce precise, well-structured written evaluations Comfort with ambiguity - real model failures are messy and require judgment, not just rule-following WHAT SUCCESS LOOKS LIKE Direct experience contributing to the training pipelines of the world's top AI labs - an increasingly rare and valuable credential in any technical or research career Deep insider knowledge of how RLHF, RLAIF, and post-training evaluation actually work at frontier labs Academic and professional credential as a Sovrano AI Expert Fellow Experience directly applicable to careers in AI research, ML engineering, and data science Earnings of €18–24/hr, paid weekly CONTRACT & PAYMENT TERMS
Internship format: 8–16 weeks, 5–15 hours per week. University credit eligibility varies by institution. €18–24/hr paid weekly. Internship certificate provided upon completion.