<h4><strong>Position:</strong> Agent Function Calling (Remote)</h4><h4><strong>Employment Type:</strong> Contractor</h4><h4><strong>Experience Required:</strong> 3+ years</h4><h4><strong>Engagement Length:</strong> 6 Weeks <em>(5 days working)</em></h4><h4><br /></h4><h4><strong>Role Overview</strong></h4><h4>This position is within a project with <strong>one of the foundational LLM companies</strong>. The goal is to assist these foundational LLM companies in enhancing their <strong>Large Language Models (LLMs)</strong>.</h4><h4>One way we help these companies improve their models is by providing them with <strong>high-quality proprietary data</strong>. This data serves two main purposes:</h4><h4>As a basis for <strong>fine-tuning their models</strong></h4><h4>As an <strong>evaluation set to benchmark</strong> the performance of their models or competitor models</h4><h4>For example, in the case of <strong>Agent Completion (AC) data generation</strong>, your task will be to simulate <strong>high-quality multi-turn conversations</strong> between a user and a smart assistant that utilizes <strong>function-calling tools</strong> to accomplish user goals.</h4><h4>You will craft these dialogues by <strong>playing both the assistant and the user</strong>, while simulating tool use where necessary to guide the assistant through <strong>complex decision-making and real-world reasoning scenarios</strong>.</h4><h4><br /></h4><h4><strong>What Does the Day-to-Day Look Like</strong></h4><h4>Design <strong>multi-turn conversations</strong> that simulate real interactions between users and AI assistants using apps like <strong>Calendar, email, maps, and Drive</strong></h4><h4>Emulate both the <strong>user and the assistant</strong>, including the assistant's <strong>tool calls</strong> <em>(only when corrections are needed)</em></h4><h4>Carefully select <strong>when and how the assistant uses available tools</strong>, ensuring logical flow and proper usage of <strong>function calls</strong></h4><h4>Craft dialogues that demonstrate <strong>natural language, intelligent behavior, and contextual understanding across multiple turns</strong></h4><h4>Generate examples that showcase the assistant’s ability to:</h4><h4>Gracefully <strong>complete feasible tasks</strong></h4><h4><strong>Recognize infeasible tasks</strong></h4><h4>Maintain <strong>engaging general chat when tools aren’t required</strong></h4><h4>Ensure all conversations adhere to <strong>defined formatting and quality guidelines</strong>, using an internal playbook</h4><h4>Iterate on conversation examples based on <strong>feedback</strong> to continuously improve <strong>realism, clarity, and training value</strong></h4><h4>Collaborate with <strong>peers and reviewers</strong> to maintain <strong>consistency and high standards</strong> in deliverables</h4><h4><br /></h4><h4><strong>Requirements</strong></h4><h4>Strong <strong>general technical reasoning skills</strong> and the ability to model real-world assistant behavior using <strong>tool-based APIs</strong></h4><h4>Ability to <strong>break down complex tasks</strong> and simulate realistic dialogues that reflect <strong>user expectations and assistant limitations</strong></h4><h4>Experience in <strong>any programming language or tech stack</strong> is acceptable</h4><h4>Strong understanding of <strong>APIs, data formats (e.g., JSON), and logical thinking</strong></h4><h4><strong>Excellent written communication skills in English</strong>, with clarity, tone, and instructional coherence</h4><h4><strong>Creativity and attention to detail</strong> in crafting realistic scenarios and responses</h4><h4>Experience working with <strong>LLMs, virtual assistants, or function-calling frameworks</strong> <em>(preferred)</em></h4><h4>Ability to <strong>follow detailed guidelines and formatting standards with high consistency</strong></h4>