Thread by @paulabartabajo_

Every aspiring data scientist I talk to is overwhelmed by the colossal amount of online courses to choose from 🤯

My solution to this problem ↓

Learning is about connecting the dots.

However, it feels like there are too many dots to connect when learning data science.

Too many courses...
Too many blog posts...
Too many technologies...

Solution: You need to change the way you learn.

As a professional data scientist, you are expected to be a problem-solver for the company or institution you work for.

You need to be good at building data science products that solve business problems.

And for that, you don't need to be an expert in Python, for example.

Instead, you need to know *enough Python* to build your solution.

And to know how much Python is *enough* you must start your learning from the end goal.

Do not learn, and then start.
Start, and then learn.

Pick a project you are interested in.

For example, if you are into computer vision, you can set yourself this goal:

"I want to build a REST API that does face recognition".

Starting from the end goal puts your mind in the "problem-solving" mode.

So you start asking yourself the right questions.

And you start googling.

#question 1: "Is there a public dataset with human faces I can use to build my model".

And you happen to find a blog post with a list of relevant datasets of human faces. You investigate a bit further and find the one that seems the best for your use case.

Boom.

#question 2: "What ML model is good for face recognition?"

And you find 5 tutorials on Youtube that covers the model-building phase.

2 minutes later you pick the one that uses PyTorch, because you are more interested in PyTorch than Tensorflow.

Boom.

#question 3: "How can I build a REST API?"

And you discover Python Flask, an easy-to-use library that is designed just for that.

You find a quick-start tutorial that shares a basic skeleton for a REST API in Flask.

Boom.

#question 4: "How do I make my API accessible to anyone on the Internet?"

And you discover a few of the most popular ways to deploy ML APIs, from beginner-friendly options like Streamlit to more advanced solutions like AWS API Gateway.

So you pick Heroku, a good middle ground.

To build this project you will need *just enough Python*, to

→ train your ML model
→ build a REST API

And you don't need to become a Python expert to do these things.

When you learn by building projects, you get 2 things:

1 → Become a problem-solver. And this is what separates a senior from a junior data scientist.

2 → Build a portfolio. Every project you make is another valuable asset you will have when looking for jobs.

In conclusion,

→ You learn by solving a specific problem.

→ When you start from the end goal, you *think* and you learn just enough of each technology that lets you build your solution.

→ As a bonus, you build a portfolio that will greatly help you when looking for jobs.

Wanna learn to build production-grade ML products?
→ 𝗙𝗼𝗹𝗹𝗼𝘄 𝗺𝗲 @paulabartabajo_
→ Join my email list: datamachines.xyz/subscribe/

Wanna help?
→ Like/Retweet the first tweet below to spread the wisdom ↓↓↓

https://twitter.com/paulabartabajo_/status/1654440610856591360

Ready to take your ML career to the next level?

Learn with me how to build a complete ML app, in the Real-World ML Tutorial and Community. Because THIS IS what companies look for.

Use the discount code "NINJA" to get a 20% discount and LIFETIME access
realworldmachinelearning.carrd.co