Chapter 1: Where You Are Now — The ML Lifecycle

In this chapter

• Vivaan from Noida — graduating from Senior 1 to a first real model.
• The complete 6-step ML lifecycle: data → model → train → eval → deploy → monitor.
• Why deployed AI is different from textbook AI.
• The 3 tools you need on your laptop: Python, scikit-learn, Streamlit.
• Your first portfolio project goal — set this week.

The 6-step ML lifecycle

Every chapter of this book teaches one of these six steps. (1) DATA — find or collect a dataset relevant to your problem: CSV, JSON, image folder. (2) MODEL — pick an algorithm (linear, tree, neural net); choose by problem, not by hype. (3) TRAIN — show examples and adjust weights; most of your time is here. (4) EVALUATE — test on data the model has never seen; accuracy alone is not enough. (5) DEPLOY — put your model behind a web page or API so others can use it. (6) MONITOR — watch how it behaves in real life; real data drifts and models age.

💻 CODE LAB Hello scikit-learn — 5 Lines to Real ML

# Install once: pip install scikit-learn
from sklearn.datasets import load_iris
from sklearn.tree import DecisionTreeClassifier

data = load_iris()
model = DecisionTreeClassifier()
model.fit(data.data, data.target)
print('Trained! Accuracy:', model.score(data.data, data.target))
# Output: Trained! Accuracy: 1.0  <- over-fit warning! Read Ch 6.

⚙ TRY IT TODAY

Go to colab.research.google.com (free, no install). Paste the code above. Hit run — you just trained your first real ML model. Now CHANGE one thing: print model.predict(data.data[:3]) to see actual predictions. Hint: predict returns flower type as 0, 1, or 2 — map them to ‘setosa’, ‘versicolor’, ‘virginica’ using data.target_names.

Questions — Chapter 1

A. Fill in the blanks

1. The 6 steps of the ML lifecycle are: Data, Model, Train, , Deploy, Monitor.
2. The free cloud Python notebook used in this book is .
3. The Python ML library taught in this book is .
4. The free deployment platform for AI web apps is .
5. Vivaan is from the city of .

B. Short answers

1. In your own words, what does ‘deploy’ mean in ML?
2. Why is accuracy alone not enough to judge a model?
3. Name 3 real Indian AIs that follow the 6-step lifecycle.
4. Write 5 lines of scikit-learn code that loads any dataset and trains a model.
5. What is YOUR portfolio project idea? Write 2 sentences.

Check your understanding — Chapter 1

Tap an answer to check it instantly. These are for practice — nothing is saved or sent.