If you wanna know can you do it, and how, stay glued for a minute. You won't regret it. So, let’s get to the point.
what does it take?
You just need a little aptitude for this; a problem solving mindset, willingness for learning and a few certifications. And rest assured, you absolutely don’t need any tech background, or a CS/AI degree.
This meme here might ease you a bit. You just have to type import keras, and keras will do all the complicated stuff for you. You just have to know what's going on.
The 50 days roadmap
A refined, organized version of my hit and trial journey to ML, albeit within 50 days
Assuming you took your maths lessons in high school seriously, and you know a little about linear algebra, matrices and a little statistics and probability, we can move to day 1 and start with the real thing right away.
If you are unsure of your maths expertise, check out my blog about maths requirement for AI/ML for info and resources. Don't fret! A zero day, or a week, would be good enough to get you going.
Day 1:
Machine Learning Specialization by Andrew Ng
A 5 hours course available on Deeplearning.ai YouTube channel and Coursera.
Day 2–3:
Python. Basic understanding and practice. Go through everything, from PRINT to FUNCTIONS and CLASSES.
From any source; Realpython, geeksforgeeks, w3schools, freecodecamp or any video tutorials like codewithmosh. Don't miss out on hands on practice and some practice exercises of every topic from Pynative or some source of your liking.
Day 4:
Numpy. Understanding and practice. With special focus on dimensions and indexing.
You should be able to code a little, and understand the syntax, so you are comfortable when it is used in assignments in the Deep learning course.
Sources: Official documentation, or same as Python. Practice exercises are a must.
Day 5:
Pandas.
Sources same as above. You don't need mastery in Python, Pandas, Numpy or Matplotlib. Basic understanding with some practice is sufficient.
Day 6:
Same as above.
If you have decided to give ML/DL/AI a go, do check the blog “Some important footnotes to 50 days to ML” before practically starting the roadmap.
Day 7–34:
(Day 15: Tensorflow, check below this)
Deep learning specialization by Andrew Ng.
Available on Coursera. It is a 5-course package, and arguably the best and most comprehensive deep learning course out there.
It can be completed in less than 3 weeks, but allocating 4 weeks to comfortably do it. You'd need breaks when you get frustrated. Let's be realistic, hmm? 😉 It's a little math-heavy, but you don't need to solve those equations yourself in practical AI, or rote learn the math. It's just the understanding of what goes on behind the scenes. So don't stress much about maths. Focus on the DL concepts, and the assignments, especially those of course 4 and 5, as you'd be using them only, not the earlier ones.
Day 15: (or whenever course 3 ends)
Tensor flow. After course 3. You'd need tensorflow in course 4 and 5. But you can't understand it before course 3 either.
A 4 video series named “Introduction to Machine Learning: ML zero to hero” on Tensorflow official YouTube channel, includes 4 notebooks for practice.
Day 35:
Download, Install, Read the documentation, Practice.
The package and documentation are all available on the official site.
Day 36:
Pycharm. Or any other Python IDE.
From official site. Download the community version, it's free. Focus on learning and using the hotkeys, as it speeds up your coding. Learning/tutorials are within the IDE. There is no need to read elsewhere.
Day 37:
Git.
From introduction to setting up and managing a git repo.
Official sites are good enough sources.
Day 38:
Flask. (Or Django). From knowing about APIs to practical usage.
Miguel Grinberg's blog is the best for Flask.
Otherwise from the official site or any source of your liking.
Day 39:
Docker. From understanding to deployment of flask APIs using Docker.
From official site.
Day 40–45:
Practice Python and Numpy in pycharm or your favourite IDE, focusing on your weak areas. By this time you'd be way more comfortable with coding than your first week.
Day 46–49:
Quick revision of Machine learning and Deep learning concepts, as you wouldn’t be going through the theory much. It's time for some practical code whacking.
Day 50:
Start building your first Deep Learning project! Or Artificial intelligence project. Whatever sounds cool to you. 😎
Congratulations 🎉 You’ve done it. The boring part is over. It's time to roll. You have come a long way. Look back and recall the time you read this blog for the first time. You didn't know a thing. And you are working on your first Machine Learning project today. Best of luck!
And if you need a roadmap for the next 50 days, check out that blog. Cheers 🥂
Don’t hesitate to hit me up for queries. But do try to search things up yourself first, it’d make you self-sufficient, a must have skill. Good luck! If you have come so far, let's connect and grow together. Here's my LinkedIn!