End 2 End Machine Learning Project

Deepak Mishra

13 min read

·

Aug 18, 2026

--

Listen

Share

← Part 1 — ML Interview Cheatsheet || Part 3 — Classification →

Press enter or click to view image in full size

1.The most important lesson

“Which model should I use?”

“What problem are we solving, how will success be measured, what data-generating process produces the data, what constraints exist, and how will the model operate reliably in production?”

2. Step 1 — Look at the big picture

What is the business problem?

In the example:

Predict median housing prices for districts.

But the existing process is manual and expensive, and estimates can be significantly inaccurate.

This is important.

The ML problem should come after understanding the business problem.

3. Frame the ML problem

You need to answer:

Is this supervised or unsupervised?
        ↓
Classification or regression?
        ↓
Batch or online?
        ↓
What is the target?
        ↓
What are the inputs?

For the housing example:

Supervised
   ↓
Regression
   ↓
Multivariate
   ↓
Batch learning