Build and ship Own a project end to end: understand the problem, build a working version, get it in front of users, improve it based on what they say. Write code that other people can read and maintain. AI-generated or not, it carries your name. Demo your work regularly to the product and engineering team, as well as to the business teams who requested it. Work with the businesses Talk to the people whose jobs your tool touches. Watch how they actually work rather than taking a feature request at face value. Translate what you hear into something you can build this week, not next quarter. AI-augmented development Use agentic AI tools (such as Claude Code) as your default workflow. Develop judgment about where these tools help and where they quietly produce garbage. Bring that back to the engineering team. Try new tools as they ship and tell us what you found, including the failures. 
