Remote LLM Trainer

About the Role

We are seeking candidates with experience in any programming language to validate and ensure the quality of training data used in developing agentic AI systems. This fully remote role involves working with function-calling methods to create structured tasks, ensuring data workflows remain consistent, accurate, and aligned with cutting-edge AI requirements.

Day-to-Day Responsibilities

• Validate large-scale training datasets and maintain strict quality standards for agentic AI use cases.
• Create, manipulate, and structure JSON tasks specifically designed for function-calling workflows (e.g., in Python/JavaScript environments).
• Work closely with AI researchers to align datasets with advanced LLM capabilities like reasoning and coding.
• Monitor data accuracy, consistency, and integrity across projects involving frontier model training.
• Contribute to documentation and best practices for dataset preparation.

Key Requirements

• Programming: Strong proficiency in Python, JavaScript, Java, or other languages for data handling and automation.
• Data Skills: Solid understanding of SQL for querying datasets; Expertise in JSON manipulation is essential.
• Focus: Exceptional attention to detail with a focus on data quality, consistency, and integrity.
• Domain Knowledge: Interest or exposure to agentic AI, LLMs, NLP best practices, prompt engineering, and RLHF concepts (Bonus).

Work Model & Offer Details

This is a fully remote contractor assignment. You will work on cutting-edge projects with leading global AI companies.

• Commitments: At least 4 hours per day, minimum 20 hours/week (options for 20h, 30h, or 40h available), with a required overlap of 4 hours with PST timezone.
• Type: Contractor assignment (no medical/paid leave benefits included).
• Duration: Initial contract of 1 month, adjustable based on engagement performance.