CampusX, led by Nitish Singh, has built one of the most respected free machine learning resources on YouTube, and the 100 Days of Machine Learning playlist is its centrepiece. What sets it apart from the endless supply of surface level tutorials is that it refuses to skip the hard bits. A typical episode will take a topic, spend real time on the mathematical intuition behind it, and only then move into implementation with scikit learn on an actual dataset, so you finish understanding not just how to call a function but why the algorithm behaves the way it does. That combination, patient theory followed by hands on code, is exactly what most self learners need and rarely get for free.
The community around it helps too, with shared notes and the code available so you can follow along rather than frantically pausing to copy from the screen. It has understandably become a rite of passage for a lot of people breaking into data science, especially across India where CampusX is something of an institution. The honest caveats are practical rather than about quality. The teaching is delivered largely in Hindi and Hinglish, which is a genuine strength for its primary audience and a real barrier for anyone who does not follow those languages, and it is only fair to say so plainly.
Being a YouTube playlist, it also comes with no formal structure, no graded assessment and no certificate, so nothing external keeps you accountable and nothing at the end proves you did it. The pacing wanders a little across a hundred plus videos, and some entries clearly assume you have watched what came before, so dipping in at random can leave you lost. None of that is a criticism of the content, which is excellent, but it does shape who this suits. My take is that for a motivated learner who can follow the language, this is a serious, genuinely deep machine learning education hiding in plain sight for free, and the main thing standing between you and the value in it is your own consistency.