Remote LLM Trainer – Agent Function Calls

About the role

We are looking for a Remote LLM Trainer to create high‑quality multi‑turn conversations that simulate real interactions between users and AI assistants equipped with function‑calling tools. Your work will directly support foundational LLM companies by providing proprietary data for fine‑tuning and evaluation.

Key responsibilities

• Design multi‑turn dialogues that emulate realistic user‑assistant exchanges using tools such as calendar, email, maps, and drive.
• Play both the user and the assistant, inserting tool calls only when corrections or specific actions are required.
• Select appropriate moments for tool usage, ensuring logical flow and correct function‑call syntax.
• Produce conversations that showcase natural language, intelligent behavior, and contextual understanding across several turns.
• Follow strict formatting and quality guidelines outlined in an internal playbook.
• Iterate on examples based on reviewer feedback to improve realism and clarity.
• Collaborate with peers to maintain consistency and high standards.

Required profile

• Strong technical reasoning and the ability to model real‑world assistant behavior using tool‑based APIs.
• Capability to break down complex tasks into realistic dialogue steps.
• Experience with any programming language and a solid grasp of APIs and data formats such as JSON.
• Excellent written English with focus on clarity, tone, and instructional coherence.

Required skills

• Programming (any language)
• API usage
• JSON data handling

What we offer

• Opportunity to contribute to cutting‑edge AI model development.
• Fully remote work environment.
• Collaboration with leading AI research teams.