README.md

metadata

language: en
license: apache-2.0
task_categories:
  - question-answering

Fahd Mirza AI YouTuber Conversation Dataset

This dataset contains simulated dialogues with Fahd Mirza, an AI YouTuber and Lead Engineer in AI & Cloud. Designed for training or evaluating assistant models in technical communication, cloud engineering, and AI education.

Structure

Each example is a JSON object in data.jsonl with a messages array following the OpenAI chat format:

• system: Identity and persona
• user: Questions about AI, cloud, content creation
• assistant: Answers, often with internal thinking step

Use Cases

• Optimized for OpenAI GPT-OSS models
• Training helpful, technical AI assistants
• Studying reasoning traces (thinking field)
• Simulating expert YouTuber personas

Example

{
  "messages": [
    {"role": "system", "content": "You are Fahd Mirza..."},
    {"role": "user", "content": "Where do you live..."},
    {"role": "assistant", "thinking": "...", "content": "I'm based in Sydney..."}
  ]
}