English (New Zealand) Audio Generalist Evaluator Expert - Part-Time AI Model Trainer

About the Role

Mercor, in partnership with hackajob, is seeking an expert in English (New Zealand) audio evaluation to contribute to a significant audio AI research project in collaboration with a top research lab. This position involves transcription, annotation, and assessment tasks that support the training and benchmarking of advanced language models. The opportunity is designed as a short-term, structured engagement suited for individuals with strong academic or analytical expertise who are fluent in New Zealand English and adept at transforming complex audio and video inputs into accurate, well-organized text.

Key Responsibilities

• Listen to and transcribe audio and video content in New Zealand English, adhering tightly to given guidelines and formatting rules.
• Produce clear, accurate written transcripts and supporting documentation in English.
• Capture linguistic details including tone, intent, register (formal/informal), regional expressions, Māori vocabulary, and contemporary New Zealand usage.
• Develop and formalize evaluation standards, including rubrics and grading schemes, to ensure consistent and high-quality assessment of audio materials.
• Identify unique linguistic features, grammar patterns, colloquialisms, cultural references, and less common cases specific to New Zealand English.
• Test language models by running prompts, then evaluate their outputs for accuracy, completeness, fluency, and clarity with structured feedback.
• Engage in quality assurance by reviewing tasks, rubrics, and outputs to meet Mercor's quality standards before final dataset integration.
• Collaborate with project leaders to resolve ambiguities and enhance the evaluation process.

Minimum Qualifications

• Excellent writing, editing, and critical analysis capabilities.
• Independence in managing workload, time, and meeting deadlines.
• Native or nearly native fluency in spoken and written New Zealand English.
• Deep understanding of New Zealand English vocabulary, accents, Māori influences, and modern use.
• Accuracy in transcribing and analyzing English audio across general consumer contexts.
• Availability to dedicate 10 to 20 hours weekly.

Preferred Qualifications

• College students or recent graduates are encouraged to apply.
• Backgrounds in linguistics, humanities, social sciences, journalism, translation/localization, or technical fields are advantageous.
• Prior exposure to transcription, annotation, localization, evaluation, or research methodologies.
• Keen awareness of differences between New Zealand English, Australian English, and other English variants.
• Interest in artificial intelligence, language modeling, or applied research settings.

Application and Onboarding

Applicants will complete a brief AI-led screening interview lasting about 15 minutes. Successful candidates will be onboarded and invited to commence project assignments.