It's all about skills, not syntax, when AI writes the code.

The path from Junior to Senior developer used to run through years of hands-on code review, pairing, and slowly earned architectural judgment. AI coding tools have compressed and scrambled that path: junior developers can now produce working code faster than ever, but deeper skills (software design, system architecture, and knowing why a solution is right, not just that it runs) are harder to absorb when AI can paper over the gaps.

This talk argues that the answer isn't to mentor less, or the same way, but to fundamentally change how we mentor within a highly functional team. Drawing on real team experience shifting from a single Senior developer providing daily one-on-one pairing to a tiered model: Senior developers deepening Mid-level developers' software design and architecture skills, while Mid-level developers in turn coach Juniors on effective, critical AI usage; we will look at what can break when you try to scale mentorship this way, what a Senior developer's job actually becomes, and how existing training content and internal curricula can be repurposed to support each tier.

Attendees will leave with a practical framework for restructuring mentorship in their own teams, and a clearer sense of what "Senior" should mean when AI can generate the boring parts of the job.