LLM Conversation Designer — Function-Call Training

Hire Feed; United Arab Emirates; On-site; AED 180,000 - 320,000; Full time

Job summary

Hire Feed is seeking an LLM Trainer - Agent Function call to enrich proprietary data for fine-tuning and evaluation of large language models. The role focuses on generating high-quality conversational data through realistic, multi-turn scenarios.

You will model both user and assistant interactions, including when to invoke function calls, to demonstrate natural language understanding and tool usage across diverse contexts, aiming to improve AI performance and reliability.

Qualifications

• Experience in multi-turn dialogue design and natural language simulation.
• Ability to emulate AI assistant behavior with tool calls.
• Familiarity with formatting guidelines and quality standards for training data.
• Strong attention to detail in crafting realistic interactions.

Responsibilities

• Design multi-turn conversations simulating real interactions between users and AI assistants using tools like calendar, email, maps, and drive.
• Emulate both user and assistant roles, including assistant tool calls when corrections are required.
• Select when and how assistants use available tools to ensure logical flow and proper function call usage.
• Craft dialogues demonstrating natural language, intelligent behavior, and contextual understanding across multiple turns.
• Iterate on conversation examples based on feedback to improve realism, clarity, and training value.

Skills

Dialogue design

NLU/NLP simulation

Tool use with function calls

Attention to detail

Job description

Hire Feed is seeking an LLM Trainer - Agent Function call to enrich proprietary data for fine-tuning and evaluation of large language models. The role focuses on generating high-quality conversational data through realistic, multi-turn scenarios.

You will model both user and assistant interactions, including when to invoke function calls, to demonstrate natural language understanding and tool usage across diverse contexts, aiming to improve AI performance and reliability.