Hello World!

This week hit me with a major self-realization. I took a step back and asked myself: With the technical knowledge I've gained so far, is it enough to make me stand out from the rest of the engineers and devs in the world? Before trying to compete with the whole world, I looked around me to understand what actually differentiates us and what makes a candidate less appealing to recruiters. That’s when it hit me: I need to seriously work on my communication skills. I can communicate properly, and being an introvert makes the hesitation make sense, but the industry won't take that as an excuse. I have to address this problem before it gets too late.

I also noticed another gap in my journey so far. I am building things and learning a lot, but I am not properly organizing it. I haven't been writing README files for my code, and there's no proper documentation for someone who wants to pick up one of my notebooks to practice or evaluate it. This is something I should have been doing from the start. Maybe I just don't like interacting with humans (haha, just kidding!), or maybe it's just because I've never done this kind of thing before. But I want to start.

To work on my communication, I’m planning to record myself teaching the concepts I learn or explaining my failures after tackling a hard topic. I'm aiming for about two videos a month to ensure a smooth start so I don't burn out and quit midway.

One Month Down!

Congratulations to myself: You did good, Ash. You maintained the streak! It's been a month, and I have covered some important foundational stuff. I haven't mastered it all yet, but I finally feel comfortable with these ML algorithms. There is definitely harder stuff yet to come, but I know I can do it.

The "Modern Algorithm" Reality Check

This past week, I was feeling a bit exhausted after wrapping up Logistic Regression. Out of curiosity, I asked an AI to list all the ML algorithms out there and tell me which ones are actually used in modern software applications.

To my disappointment, the top three (like Gradient Boosted Trees and other Tree-based algorithms) were nothing I had learned so far! For a second, I felt like, If this is what the industry uses, why not just learn those directly? But at the start of this journey, I made a promise to myself to follow one roadmap and one resource path. So, I don't care—I am sticking to the CampusX roadmap for now to build my foundation.

The Algorithm Showdown

Technically, last week was all about understanding the mathematical intuition behind algorithms like KNN, Naive Bayes, SVM, and Decision Trees.

To see how they stack up, I implemented the same dataset across all of them to see which performed better.

Here are the top results:

• Winner: SVM (Support Vector Machine) with an 84% accuracy.

• Runner-up: Logistic Regression at 80%.

And I guess that's all I did for the past week. I'm really struggling to find time to do more because of a hectic college schedule. Attending mindless lectures that provide no value leaves me feeling completely exhausted by the time I'm free. But I know crying about it isn't going to do me any good.

So, see you in the next one with something optimistic.

Till then,
DO Some Code.