This presentation is a roadmap for running a successful Data Science and AI team. It advocates for a focus on experimental agility, and what is needed to achieve that from different angles. We'll explore the importance of teamwork and a culture of continuous learning. We'll talk about the importance of testability, reproducibility, and learning from mistakes. Also, we'll discuss sharing results and insights within the team to foster collective learning. Lastly, we'll circle back to data quality, emphasizing its crucial role in model performance.