How does ML education feel today

I came across a study by J'son & Partners on AI workforce training in the USA, China, and Russia. The authors look at how countries are trying to bridge the gap between the demand for ML specialists and what the educational system produces. Demand is growing everywhere, and the key challenge is common — the competencies of graduates do not match the current business tasks. The approaches are different. China focuses on centralization: the government sets unified standards, and from 2025, AI will be mandatory in schools from the age of 6, all living on a single state platform. It scales well, but for non-standard tasks, specialists are often sought abroad. In the USA, the university model prevails — universities compete with each other and update programs according to the market. Education is expensive, and big tech companies also participate: Google, Meta, and others launch joint programs where one can pursue a master's or PhD while working at the company. In Russia, the model is based on cooperation between big tech companies and universities: companies launch joint programs, open basic departments, and run their own schools like ShAD. As a graduate, I can say that this format provides both an academic foundation and an understanding of how ML works in production. The latter is difficult to obtain in a pure university setting — this knowledge lives within teams and products. So if you are choosing an ML program, look not only at the university but also at which companies are behind the training. It’s great if you can join a program where you can write a PhD or master's thesis directly in the company. Many of my acquaintances did this in programs with Facebook AI Research.