The posting
ABOUT THE ROLE
Join an early-stage research team at a European AI data company and investigate whether expert-generated data improves language model performance. You will conduct hands-on research, build evaluation infrastructure, and help determine which research and data initiatives to pursue.
WHAT YOU'LL DO
- Post-train language models on expert datasets and measure their impact, investigating when results fall short.
- Design and run evaluations of frontier language models in expert domains.
- Select and execute suitable training methods, including supervised fine-tuning, reinforcement learning, and preference tuning.
- Translate research findings into dataset specifications and recommendations for future work.
- Publish research papers, benchmarks, and written research updates.
- Coordinate research findings with data, client, partnership, and marketing teams.
WHAT WE'RE LOOKING FOR
- Hands-on research experience with reinforcement learning environments for language models or RL-based post-training, such as RLHF, RLVR, reward modeling, or preference optimization.
- A completed or nearly completed PhD in machine learning or NLP, or an equivalent research record demonstrated through publications or AI lab and industry research work.
- Experience personally post-training and evaluating language models end to end, and building evaluations or benchmarks.
- Research publications at leading machine learning or NLP venues, or experience building evaluation tools.
- Early-career research profile, from near PhD completion to a few years post-PhD.
COMPENSATION & BENEFITS
Competitive salary and equity. Relocation support is available. Visa sponsorship details were not provided.
LOCATION
Full-time, on-site in Barcelona, Catalonia, Spain. Relocation support is available.
