Things I Wish I Knew Before Learning Machine Learning

I spent three months grinding through ML tutorials. Here’s what I’d tell myself at the start.

1. Math first, code second

You can copy-paste model code without understanding it. But when it breaks, you’re lost. Spend time on linear algebra and probability basics — it pays off later.

2. Start small

Don’t start with GPT or diffusion models. Start with linear regression on a CSV file. Understand what a loss function actually means.

3. Sklearn is your best friend

Before touching PyTorch or TensorFlow, get fluent in scikit-learn. It teaches the concepts cleanly.

from sklearn.linear_model import LinearRegression
model = LinearRegression()
model.fit(X_train, y_train)
print(model.score(X_test, y_test))

4. Overfitting is real

Your model will perform great on training data and terribly on new data. Learn about train/test splits, cross-validation, and regularization early.

5. Data prep is 80% of the work

Cleaning data, handling missing values, feature engineering — this is where most time goes. Embrace it.