LLM Multi-Turn Conversation Designer & Tool-Use Simulation Specialist
I designed high-quality multi-turn conversations for training and evaluating AI assistants. This work included creating realistic dialogues between users and AI systems, modeling natural communication patterns, and simulating rich interactions across productivity applications such as calendar, email, maps, and cloud drive services. I authored both user messages and assistant responses, including accurate tool-use behaviors and corrective actions when the assistant needed to adjust its approach. Each conversation demonstrated contextual understanding, step-by-step reasoning, and smooth handling of both feasible and infeasible requests. The goal was to improve LLM performance in task completion, tool orchestration, and general conversation quality. I ensured clarity, alignment with internal playbooks, and high-quality standards through iterative revisions and collaboration with reviewers.