Duke's AIPI program is a Master of Engineering, not a Master of Science, and that tells you what it is for. It is aimed at engineers and technical people who want to build AI products, so the curriculum puts applied machine learning next to product development and business. Searches for it are small, about 70 a month in the US, but they have more than doubled in the last year. The degree is 30 credits. The technical core covers sourcing data for analytics (AIPI 510), modelling process and algorithms (AIPI 520), deep learning applications (AIPI 540), legal, societal and ethical implications of AI (AIPI 560) and operationalizing AI, which is MLOps (AIPI 561).
Around that sit business fundamentals and management of high-tech industries, a career strategy course, an industry seminar series, an internship or industry project, and an industry capstone with an external sponsor. Before you start you also take an online Python and data science math bootcamp. I like the inclusion of MLOps as a core course. Most programs still treat deployment as an afterthought, and it is where a lot of real AI work actually happens. You can do it on campus in an accelerated 12 months or a standard 16 months, or online part-time over 24 months with two required on-campus residencies.
For the online version, a Duke partner page lists 2025-26 tuition at $28,920 a year for two courses a semester, which works out to roughly $58,000 for the degree before travel to the residencies. Duke's own page did not show me a current figure, so check before you budget. There is also a cheaper way in. The AI Foundations for Product Innovation graduate certificate is four online courses (510, 520, 540 and one technical elective) over about 15 months, paid per course. It requires an undergraduate degree in science or engineering, or equivalent technical experience, plus two semesters of calculus.
Finishing it lets you apply to the full MEng, although Duke's page does not promise that the credits carry over automatically. My honest take: the course design is good, and if I were hiring an AI product engineer I would be happy to see this degree. But the cost comparison is hard to escape. Georgia Tech and UT Austin will give you a more rigorous technical master's for around a sixth of the price. What Duke adds is a smaller cohort, more structured career support, the product and business courses, and the Duke name.
For some people, especially with employer funding, that is worth it. For most self-funded learners, I would rather they took a cheaper degree and learned product skills on the job.