AI Specialization Program

🔷 WEEK 1  Lesson 4

Title: Python for AI Systems — Writing Clean, ML-Ready Code

Lesson Objective

By the end of this lesson, learners will:

Refresh core Python skills needed for AI

Understand functions & modular coding

Write structured, reusable code

Understand NumPy fundamentals

Write ML-ready, clean engineering code

This lesson is about discipline.

AI engineers do not write messy notebooks.

They write structured systems.

1.  Python Refresher — What Actually Matters for AI

You do not need all of Python.

You need:

Variables

Data types (int, float, string, bool)

Lists

Dictionaries

Loops

Conditionals

Functions

Imports

AI Engineering requires clarity, not clever tricks.

Example mindset:

Bad:

x=[1,2,3];y=[4,5,6];print([x[i]+y[i] for i in range(3)])

Engineer mindset:

def add_vectors(a, b):

    result = []

    for i in range(len(a)):

        result.append(a[i] + b[i])

    return result

Readable. Maintainable. Debuggable.

2. Functions & Modular Coding

Functions are critical in AI systems.

Why?

Because AI projects grow quickly.

Instead of writing everything in one file, engineers:

Separate data loading

Separate preprocessing

Separate model training

Separate evaluation

Example structure:

project/

│

├── data_loader.py

├── preprocess.py

├── train.py

├── evaluate.py

└── main.py

Each file has clear responsibility.

This is professional coding discipline.

3. Writing ML-Ready Code

ML-ready code is:

Clean

Reusable

Well-named

Structured

Commented where necessary

Example:

Bad variable name:

x = data[:,0]

Better:

feature_column = data[:, 0]

Engineers optimize for readability, not shortcuts.

4. Why NumPy Is Essential

NumPy is the backbone of ML computation in Python.

It allows:

Fast numerical operations

Vectorized computation

Matrix multiplication

Linear algebra operations

Without NumPy, ML would be too slow.

5. NumPy Basics (Core Concepts Only)

Creating Arrays:

import numpy as np

vector = np.array([1, 2, 3])

matrix = np.array([[1, 2], [3, 4]])

Shape:

vector.shape

matrix.shape

Shape tells you structure:

(rows, columns)

Understanding shape prevents many ML bugs.

 6. Vectorized Operations

NumPy allows element-wise operations:

a = np.array([1, 2, 3])

b = np.array([4, 5, 6])

a + b

No loops required.

This is faster and cleaner.

7.  Matrix Multiplication in NumPy

X = np.array([[1, 2],

              [3, 4]])

weights = np.array([0.5, 1.0])

prediction = X @ weights

That single line:

X @ weights

Is what powers most ML models.

Understanding this is critical.

8. Engineering Discipline: Avoid Notebook Chaos

Common beginner mistake:

Everything in one Jupyter notebook.

Professional approach:

Prototype in notebook

Move stable code into .py files

Use functions

Create reusable pipelines

Your future self must understand your code.

 Mini Practical Exercise

Create a Python script that:

Creates a NumPy matrix (5 rows, 3 features)

Creates a weight vector (3 values)

Computes predictions using matrix multiplication

Wrap computation inside a function.

Print shape of input and output.

This builds production habit early.

Week 1 – Lesson 4 Outcome

Students now:

✔ Refresh core Python skills

✔ Understand modular coding

✔ Understand clean engineering structure

✔ Understand NumPy arrays & shapes

✔ Perform matrix multiplication in code

✔ Write ML-ready structured scripts

Next:

🔷 Lesson 5 — Data Structures & Algorithms for AI

Then

🔷 Lesson 6 — Git & GitHub Workflow + Mini Engineering Setup Project.

Powered by Soft AI Africa | Training the Next Generation of AI Leaders in Africa.