can u please make a roadmap 🙂

Good high-level summary ;)

I wouldn't really describe numpy pandas and matplotlib as "ML libraries"; more like Data Science Libraries

Thanks. Now I will be going.😊

Basics of ln and calculus aren't enough 😅

​@@francescocavalieri5443 I do think they are :) For most people at least, especially if you are talking about ML engineering. To understand the basic ML algorithms, you really don't need complex stuff. You need to know how derivatives work and how to do matrix multiplication. Those are the fundamentals. Sure there are things like cosine similarity. But generally, you just need to know how to apply it. But even understanding it is arguably not that complex. The most difficult thing will probably be Eigenvectors and Eigenvalues, but those are really only relevant for one of the common ML interview questions - questions about PCA. Of course, you can always go deeper. And at some point you will understand the math better and better. But still, from my experience, and that of the ML engineers I have spoken to, the basics really are enough.

Wow, thank you! I am genuinely really happy to hear that these videos provide value to experienced engineers like yourself!

Thanks Max! Perhaps my recent video on how I built my recent project can help you :) kzfaq.info/get/bejne/d993rcl4tteul40.html

Thank you! I really appreciate it :) You keep it up as well! 🚀

I think what makes math fun to me are the amazing things you can do with it! With a list of rules, you can go from having an idea, to formulating it to a single mathematical equation that can be used to predict the future. If you see the "power" and logic behind these math equations, they become much more fascinating and fun to learn :) At least for me!

Think about what you need to learn it and like it.

Yeah, this can be tough... It might be more challenging, but perhaps you might want to look outside of your country for opportunities, or look for remote ones.

@@borismeinardus ya thats what am trying to do right now sadly most remote internships ends up to be scams

Yeah, I see... Try building projects publicly, e.g. by posting about it on LinkedIn or doing pull requests to cool OS projects! Try joining online communities, e.g. on discord. The point is, try to not only build your skill, but also get to know people outside of your country, who you can demonstrate your skill to. Sooner or later, there will be a person who can refer you to an internship/ job or even invite you himself. I know this might seem like an annoying workaround, but if you are struggling to get internships through standard applications, you need to get there by knowing people. It will take time though! That said, knowing people in general is insanely valuable in any case! So I can recommend it even without the use case of getting an internship/ job.

You can definitely start learning as soon as in high school, yes! But it will take time (very likely several years) to really get good. The learning never stops :)

I have a video on the different ML jobs :)

@@borismeinardusshould’ve replied with that instead

Thank you! Really happy to hear that 🚀☺️

Thank you very much! ❤️❤️

Thank you! I'll create a video on advice on how to create a cool ML project and what is important for ML projects! Does that at least kind of work? :)

@@borismeinardusomg, I will watch it for sure man 🔥

It depends on the company you are applying for. But in general, especially if you are at the beginning, normal SWE internships/ jobs are a perfect place to start!

Very glad it was helpful!

Keep it up!

It really depends on where you are. If you have no idea about ML, then start with the ML basics, but if you are already somewhat into it and want to get to deep learning, this one is quite cool! introtodeeplearning.com/ There are really a lot of courses and in the end it really doesn't matter that much. You just need to keep going and you will learn many concepts when you encounter them in the wild :)

@@borismeinardus thank you very much

The journey of being AIML Engineer is a bit hard my friend, you mentioned you are from India, I suggest you to do courses on Full stack Data science and land in a job or internship first to the end of your 6th semester, and eventually you can go to ML field job, till then learn python very well, be good at math, learn communication skills and do some competitive coding as well.

Wow, that was over a year ago haha, but yeah, you are probably right. I never really gave an update on that. I was rejected :( My next video will all my experiences of past years! I hope that will provide a lot of guidance, help, and perspective :)

@@borismeinardus Thanks for the update boris, I'm sorry to hear about the rejection, but I really appreciate your openness to share your experiences....looking forward to your next video about your past years' experiences. it will be really insightful and valuable for all of us. Keep up the great work

Haha very good guess and reasoning! I don't want to share too much just yet, but it's not Citadel ;)

@@borismeinardus do you know how much longer you will wait until you reveal it? :)

probably in the coming weeks/ months 🫣

​@borismeinardus stop the CAP 🧢😀

🥹 Thank you!!!

Does this person who told you CUDA is old work on AI Ethics or explainable AI? Then he is veeeery biased haha I would definitely argue against what he recommends. AI Ethics and explainable AI are important, but to say CUDA or learning about ML hardware optimization is not relevant is very wrong. At least that is my opinion and that a fairly confident one.

​​@@borismeinardus He is into DevOps role and works around infrastructure related stuff at fintech company I want to focus more on learning infrastructure and ML components of AI during my masters, so when I messaged him on LinkedIn for guidance about electives He said CUDA programming during 2010 was very best skill to get job. It's relevant today as well, but I will need to put lot of efforts on learning it and related things. He said just like CUDA was at peak in 2010, explainable AI/interpretable ML, AI Ethics, Transfer Learning are good things to learn in 2025 He asked me to get familiar with frameworks which are based on CUDA, but easy to implement things using it as abstraction layer Anything you can suggest me? I know theoretically and practically Supervised and Unsupervised Learning and related topics I was thinking to focus on bit of DevOps and infrastructure related things for third semester and in fourth semester to focus on neural networks and deep learning

One more point he mentioned was I will need to take lot of efforts from programming perspective as the course uses C++ and CUDA And I am from non-CS background (Mechanical Engineering) with intermediate Python skills and OOP using Java

Well, it will be more challenging to learn these skills without a more extensive CS background, but that is always the case. Some topics will be easier for some people and harder for others and vice versa. But to say it is not relevant is simply unfair and false. At least in my opinion, and that of those who I have asked.

I'll do my best! 🫡

Really happy to hear that!

Hey! 😊

@@borismeinardus hay I am computer science student from India and I am learning about AI ML concept and building a project . My question is 1) To build a project what are the main concerns about the project ? 2) what are the main aspects of building a project and doing research on a particular topic . how we can build documentation as a production level project or research . 3 ) what should be the focus to solve the problem and to do research?

Those are amazing questions! You are looking for the right things, which makes you already ahead of many others. I will create a video soon where I hopefully answer all these questions :)

@@borismeinardus Thank you so much for your encouragement!😊 I'm looking forward to your video. In the meantime, do you have any resources or tips that might help me get started on addressing these questions?

😎

There are a few sources! On YT I can recommend the channels "AI explained" and "Yannic Kilcher"

Yes