Microsoft pairs junior engineers with senior mentors to learn AI-assisted development

Microsoft is putting senior engineers alongside early-career developers to teach skills that generating code alone cannot supply. In an October 15 announcement, the company said it has launched PRAISE, an internal mentoring program, and is partnering with Carnegie Mellon University on curriculum. The goal is to develop engineering judgment while juniors learn to work with AI.

The learning problem behind faster code

The announcement, written by Microsoft executives Mark Russinovich and Scott Hanselman, starts with a question about training rather than output: how does someone learn software engineering when AI handles more of the implementation? The authors argue that software engineering careers and computer science degrees still matter, even when tools can quickly produce an app.

Their reasoning rests on the work that surrounds code. They say more engineering time now goes toward design and architecture, but developers still need to understand the underlying software. Producing code with AI is only one part of building a product; engineers remain responsible for whether the finished technology works safely and effectively.

Michael Hilton, a Carnegie Mellon teaching professor and associate department head for education, describes the training challenge in the announcement. Early-career engineers previously performed many tasks that built expertise, he says. His case for the partnership is that academic research and practitioners’ experience can be brought together to address the daily realities of software development.

A training model borrowed from nursing

PRAISE stands for Preceptorship for AI-accelerated Software Engineering. A preceptor is an experienced professional who guides a newcomer through practical work. Microsoft borrowed the approach from nursing, where new nurses work with experienced colleagues to build confidence and real-world experience. Here, senior engineers take that guiding role.

The workplace structure gives junior engineers room to work through questions and problems with their mentors. Microsoft says senior engineers should recognize that newcomers may lack the confidence to contribute at an experienced colleague’s speed and scale. Building rapport is intended to make space for learning by doing, rather than simply watching someone else solve a problem.

• Senior engineers explain their approach to problems and share tips through concrete examples.
• Mentors advise juniors on where to focus, based on their existing skills and knowledge.
• Junior engineers are taught to test assumptions and discuss unfamiliar changes with colleagues.

AI changes the work; it does not remove human responsibility for outcomes.

Mark Russinovich and Scott Hanselman, Microsoft

University instruction alongside the day job

The Carnegie Mellon curriculum partnership adds formal education to that workplace guidance. Participants will take classes and join small-group discussions connecting computer science theory with practical software engineering and their work at Microsoft. The courses are intended to give them early experience deploying solutions ready for production use, not just producing code.

Microsoft describes the arrangement as a shared training responsibility: neither universities nor employers should have to carry it alone. Classroom instruction is meant to complement what engineers learn from their teams and preceptors, rather than replace that experience. The company also treats deciding when to use AI as an ongoing part of education.

The stated destination is broader than proficiency with a coding assistant. Microsoft wants participants to develop systems thinking, judgment and the confidence to build and deliver software. Its ambition is to help junior engineers grow from producing code into deciding what deserves to be built. Those are the program’s goals, not demonstrated learning outcomes.