AI Prompts Evaluator – Hindi (Paris, 25h/wk, Onsite)

TransPerfect; Paris; Hybride; EUR 30 000 - 39 000; Temps partiel

Résumé du poste

DataForce by TransPerfect is seeking a detail-oriented professional to design and evaluate prompts for AI language models in Paris. This onsite, 3-month role supports Hindi-English language tasks, including data collection, transcription, and annotation across audio and text data.

You will review model outputs, define evaluation criteria, and report insights to stakeholders, ensuring high data quality and compliance with guidelines.

Qualifications

• Idiomatic fluency in Hindi.
• Writing skills in Hindi (Devanagari script).
• Advanced level of English.
• Be in Paris (onsite position).
• Strong attention to detail and critical thinking skills.
• Basic programming knowledge and ability to review code is a plus.
• Rigorous, analytical, and methodical approach.
• Efficient execution of assigned tasks and resilience to dynamic working environment.
• Familiarity with AI language models and evaluation processes is a plus.
• Experience or strong interest in physical teleoperations, particularly robotics machine manipulation, is a plus.

Responsabilités

• Develop prompts and evaluation criteria for AI model testing.
• Review and rate model responses against guidelines and quality standards.
• Investigate and document failure cases with actionable insights.
• Communicate observations, feedback, and ratings to stakeholders.
• Apply guidelines to maintain high data quality.
• Identify and source Hindi-language audio content from online platforms.
• Transcribe Hindi audio recordings for language model training.
• Annotate and label audio/text data per project specs.
• Evaluate Speech Recognition and Text-to-Speech systems.
• Curate golden datasets for model development.
• Operate robotic equipment to perform tasks and record video sessions when needed.

Connaissances

Hindi fluency

English proficiency

Description du poste

DataForce by TransPerfect is seeking a detail-oriented professional to design and evaluate prompts for AI language models in Paris. This onsite, 3-month role supports Hindi-English language tasks, including data collection, transcription, and annotation across audio and text data.

You will review model outputs, define evaluation criteria, and report insights to stakeholders, ensuring high data quality and compliance with guidelines.