Reproduce “From Junior to Senior: Allocating Agency and Navigating Professional Growth in Agentic AI-Mediated Software Engineering” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Dana Feng, Bhada Yun, April Yi Wang

A study from ETH Zürich examined how junior and senior software engineers distribute decision-making authority when using generative and agentic AI tools and how these tools affect professional development and mentorship. The research revealed that AI can enhance efficiency but also poses risks to skill development and ownership for juniors, highlighting the persistent importance of human mentorship and the value of AI prompt histories for insights into engineer's thought processes.

Key insights

  • Organizational policies and inherent AI tool designs significantly pre-configure the distribution of agency in AI-mediated software engineering, often overriding individual choices.
  • The critical, foundational skills of senior engineers, like system-level thinking and understanding consequences, remain indispensable for effectively steering AI and ensuring robust software development.
  • Professional growth for junior engineers shifts from purely technical mastery to "earning judgment through deliberate restraint," requiring critical evaluation of AI outputs and understanding when not to delegate, while human mentorship focuses on context-specific "why-questions" and instilling ownership.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.