Jeff Haemer, an experienced software developer and instructor, sought to deepen his understanding of AI and refresh his Python skills. He enrolled in a Python Agentic AI cohort, initially setting low expectations due to his perceived gaps in knowledge. Despite encountering challenges, particularly with interface integration during a Telegram bot week, the course structure allowed him to progress. The program emphasized building the core agent logic before developing interfaces, a design that proved resilient when one interface lagged.A significant learning hurdle for Jeff was mastering the concept of mocking, which involves using substitute services for testing. He found this practice initially confusing, even after consulting with other developers. However, the course eventually made mocking essential, enabling him to distinguish between unit and integration tests and improving his understanding of Python testing frameworks like pytest. Jeff's commitment to rigorous testing resulted in a comprehensive test suite that caught actual issues.He utilized AI as a teaching assistant, seeking help with Python implementation details rather than code generation. This approach allowed him to build the application by hand, reinforcing his learning. The primary outcome for Jeff was not just the functional agent application he created, but a solid mental model of how AI agents operate. He tested this understanding against external examples, confirming his grasp of the underlying processes. This newfound clarity transformed AI from an abstract concept into a tangible area of expertise.