This fall I am participating the #mlzoomcamp held by Alexey Grigorev and his crew at DataTalks.Club. Except a few courses at the university and random readings on the internet, I didn’t have much exposure to machine learning but I wanted to get closer to it since the beginning of my career. As a coincidence, I’ve been contributing to a machine learning product at my current client in the last half a year, so I managed to gain some practical experience before jumping into the Zoomcamp.
The material is based on Alexey’s book, and has a weekly live stream session, where we discuss the homework from the previous week and further questions. The weekly learning material is transmitted on Youtube - the videos consist of theoretical and practical parts. As mentioned, we also get homeworks related to the weekly topic(s), which have to be submitted before the live sessions on Mondays.
During the first weeks we learned about the toolset of machine learning (e.g. numpy, pandas), linear and logistic regression, as well as model evaluation and deployment techniques. Currently we are getting closer to the “midterm project”, where we will have the chance to apply the things we’ve learnt in the recent weeks by implementing an end-to-end ML solution.
The course enlightened several things, which I had just vague understanding about, e.g. I really liked how Alexey explained the linear algebra behind regression using Python. I miss a bit more details about the each techniques, would be happy to learn their (dis)advantages and uses cases where they should(n’t) be used. But maybe this will come with the practice…