Microsoft and Carnegie Mellon University are pairing early-career engineers with experienced mentors and coursework designed for software development shaped by AI. The programme, called PRAISE, is Microsoft’s answer to a practical question: how do new developers build the judgement to review and own code they may not have written line by line?
Microsoft AI Watch analysis
What happened
PRAISE stands for Preceptorship for AI-accelerated Software Engineering. Microsoft says its internal programme pairs junior engineers with senior preceptors, while Carnegie Mellon is helping develop a curriculum to connect computer science theory with production software work.
Participants get space to work through problems with mentors, test assumptions and discuss unfamiliar changes. The coursework includes classes and small-group discussions; Microsoft says the aim is to teach engineers to use AI tools thoughtfully while building systems thinking, judgement and accountability.
Why it matters
AI can speed up code production, but that does not automatically teach a newcomer why a system is designed a certain way, whether generated code is sound or what deserves to be built. Microsoft’s approach puts workplace mentoring alongside university teaching, rather than treating either as a substitute for the other.
It is also a specific response to the worry that AI could shrink the entry-level work through which engineers gain experience. PRAISE is intended to change what junior developers learn, not to pretend the job has stayed the same.
Our read
The useful idea here is not that AI makes mentoring newly fashionable. It is that producing code and understanding software are different skills, and teams still need people who can take responsibility for the result. Pairing real work with experienced guidance is a sensible way to teach both.
The test will be whether the programme gives junior engineers meaningful responsibility and helps them develop sound judgement, not simply whether they become quicker at prompting. A training plan is a start; the engineers it produces are the evidence.
What to watch
- What the PRAISE curriculum teaches and how it is assessed.
- Whether Microsoft or Carnegie Mellon shares results from the programme.
- Whether similar mentoring and coursework models spread to other employers and universities.
Discussion spark: As AI takes on more coding tasks, should junior engineers learn mainly through mentored workplace projects, or should universities take the lead in teaching how to review and own AI-generated code?
Sources and evidence
- What will the next generation of developers do when AI writes the code? – Microsoft (15 October 2026, 15:14 UTC)
not affiliated with or endorsed by Microsoft