The Language Technologies Unit at BSC is currently engaged in the training of foundational multilingual Large Language Models. As we begin the instruction and alignment phase, we are seeking annotators, evaluators, and translators to join a dynamic team of enthusiasts dedicated to advancing cutting-edge AI systems.

We are looking for 20 skilled and detail-oriented professionals to analyze, annotate and create data in Spanish, Catalan, Galician and Basque. Your work will be critical in ensuring the quality and accuracy of the multilingual datasets that power our AI models.

The ideal candidate is fluent in at least two languages, has exceptional attention to detail, and is passionate about working at the intersection of language and technology.

Key Duties

Candidates will be involved in a mix of several evaluation, annotation, and content creation tasks aimed at improving AI models. These includes, among others:
Identifying undesirable behaviors and vulnerabilities in AI models, performing human evaluations of large language models, and reviewing AI-generated responses for quality.
Evaluating and labeling LLM-generated and human-generated content, including preference, instruction, and safety-related data.
Revising synthetic instruction data, creating and refining targeted prompts, and writing or optimizing prompts and responses to enhance model performance.
Transcribe audio recordings.