AI Trainer – Audio Evaluation & Multilingual Benchmarking (Luxembourgish)

About the role

Innodata is seeking freelance AI Trainers to support an upcoming research project focused on audio evaluation and multilingual benchmarking. The work involves assessing the quality of Text‑to‑Speech (TTS) and Speech‑to‑Text (STT) outputs in Luxembourgish, helping to improve language technology for AI systems.

Key responsibilities

• Evaluate audio outputs from TTS and STT systems for linguistic quality, accuracy, fluency, and naturalness.
• Review and annotate recordings, highlighting pronunciation, grammar, and dialectal variations.
• Contribute to the creation of multilingual benchmarks that reflect Luxembourgish cultural and regional nuances.
• Deliver structured feedback within defined deadlines.

Required profile

• Native or near‑native fluency in Luxembourgish.
• Strong grasp of Luxembourgish grammar, pronunciation, and dialectal differences.
• Prior experience in linguistic annotation, quality assurance, audio evaluation, or AI/LLM‑related projects (preferred).
• Ability to work independently on structured tasks and meet deadlines.

Required skills

• Familiarity with Text‑to‑Speech (TTS) and Speech‑to‑Text (STT) technologies.

What we offer

• Flexible remote work arrangement.
• Opportunity to contribute to cutting‑edge AI research.
• Collaboration with a global team of subject‑matter experts.