Part-Time AI Research Associate — MacBook Pro Required

Obsidian; Greater London; On-site; GBP 21,000 - 28,000; Part time

Job summary

Mercor is seeking detail-oriented individuals to support a research project with a leading AI lab. You will help benchmark and improve cutting-edge AI models as part of a small, focused team.

Part-time commitment of approximately 10-20 hours per week, with responsibilities including developing high-quality research questions, answers, and evaluation materials to train advanced language models. Immediate start preferred.

Qualifications

• Must have strong online research and communication skills.
• Ability to gather and clearly summarize information from diverse sources.
• Excellent written communication skills.
• Currently pursuing or recently completed Masters studies, or have 2-3 years of relevant experience + Bachelor's from a prestigious institution.
• Interdisciplinary degree is a nice to have.
• Must have a MacBook with M-series or Apple Silicon (ARM) Macs running macOS 15 or higher.

Responsibilities

• Part-time commitment of approximately 10-20 hours per week.
• Create high-quality research questions, answers, and evaluation materials to train advanced language models.
• Maintain clearly structured responsibilities without the high pressure typically found in internships or full-time roles.
• Immediate start preferred.

Skills

Strong online research

Communication skills

Summarizing information

Written communication

Interdisciplinary degree

Education

Masters studies

Bachelor's degree

Tools

MacBook with M-series

Apple Silicon/macOS 15+

Job description

Mercor is seeking detail-oriented individuals to support a research project with a leading AI lab. You will help benchmark and improve cutting-edge AI models as part of a small, focused team.

Part-time commitment of approximately 10-20 hours per week, with responsibilities including developing high-quality research questions, answers, and evaluation materials to train advanced language models. Immediate start preferred.