Python Basics for AI Development

By the end of this lesson you will be able to:

Declare variables and work with Python's core types (str, int, float, bool, list, dict)
Write functions with parameters, return values, and type hints
Use if/elif/else, for loops, and while loops confidently
Read input from the terminal and print formatted output
Run a Python script from the command line on Mac or Windows

New to this course? Each lesson ends with collapsible examples showing how the concept applies to three real projects — DocBot, HelpDeskBot, and CodeBot. Head to the Phase 0 Overview (/learn/0) to meet them before diving in. These examples are optional — the lesson teaches the concept completely on its own.

1. Why Python for AI

Python became the dominant AI language through ecosystem timing, not design. The scientific computing libraries researchers needed — NumPy, SciPy — were built in Python in the early 2000s, and every major AI framework followed. Today, the Anthropic SDK, LangChain, LangGraph, ChromaDB, scikit-learn, and PyTorch all treat Python as their primary interface.

Three reasons it matters for this course:

The ecosystem. When a new AI library ships, it ships Python first — sometimes Python only for months. If you want to use the latest tools as they arrive, you need Python.

Readability. AI systems involve enough conceptual complexity — embeddings, retrieval pipelines, agent loops — that the implementation language should add as little cognitive load as possible. Python's syntax is minimal: def for functions, for item in collection for loops, if/elif/else for conditionals. No semicolons, no type declarations, no class boilerplate just to run code.

Interactivity. Python's REPL and Jupyter notebooks let you run a single expression, see the result, modify it, run again. That tight feedback loop is how AI systems get built.

Python is slower than compiled languages like Java or C# for raw compute. For AI workloads, this rarely matters. The expensive operations — matrix multiplication, embedding lookups, LLM API calls — happen inside libraries written in C++ or CUDA. The Python code is the coordinator, not the worker. Coordinating is fast regardless of language.

2. Variables and Types

Python is dynamically typed. You do not declare the type of a variable — you assign a value and Python infers the type at runtime.

name = "Tom Fischer"     # str
employee_id = 8          # int
confidence = 0.87        # float
is_resolved = False      # bool
assigned_to = None       # NoneType — absence of a value

If you are coming from a statically typed language like Java, C#, or TypeScript, this will feel unsafe at first. The trade-off is real: Python will not catch a type error until runtime, whereas TypeScript would catch it at compile time. For AI scripts and small bots, the cost is low and the gain in brevity is high. For production services, type hints (covered below) restore some of the safety.

The five fundamental types cover almost everything in Phase 0:

str — text. Strings use single or double quotes interchangeably. Multiline strings use triple quotes.

response = "Your password reset request has been received."
multiline = """TechCorp IT HelpDesk
Type 'quit' to exit."""

int and float — numbers. int for whole numbers, float for decimals. Arithmetic works as expected. Integer division uses //, remainder uses %.

tickets_today = 12
resolution_rate = 0.73
hours_saved = tickets_today * 0.25    # → 3.0
category_id = tickets_today // 5      # → 2 (integer division)
remainder = tickets_today % 5         # → 2

bool — True or False. Python capitalises booleans: True and False, not true and false. Comparison operators return booleans: ==, !=, <, >, <=, >=.

is_urgent = True
ticket_count = 15
needs_escalation = ticket_count > 10   # → True
is_empty = ticket_count == 0           # → False

None — the absence of a value. Use None where you would use null in Java/JavaScript or null in C#. Check for None with is None, not == None.

assigned_employee = None
if assigned_employee is None:
    assigned_employee = "tom.fischer@techcorp.com"

String Formatting: f-strings

Python has several ways to format strings. Use exactly one: f-strings. They are readable, fast, and work in every context you will encounter.

name = "Tom Fischer"
ticket_id = "TKT-0042"
status = "resolved"

# f-string: prefix the string with f, wrap expressions in {}
message = f"Hello {name}, your ticket {ticket_id} is {status}."
# → "Hello Tom Fischer, your ticket TKT-0042 is resolved."

# Expressions work inside {}
count = 7
summary = f"Processed {count} tickets. {count * 0.25:.1f} hours saved."
# → "Processed 7 tickets. 1.8 hours saved."

If you are coming from a statically typed language like Java or C#, f-strings are equivalent to String.format("Hello %s", name) in Java or $"Hello {name}" in C#. In JavaScript/TypeScript they are equivalent to template literals: `Hello ${name}`. The Python syntax uses curly braces inside an f-prefixed string. Pick this pattern and never use % formatting or .format() — they are older styles you will encounter in legacy code but should not write new.

Type Hints

Python supports optional type annotations. They do not change runtime behaviour — Python ignores them during execution — but they document intent and enable static analysis tools like mypy to catch type errors before you run the code.

def get_response(message: str) -> str:
    # message is expected to be a str
    # this function is expected to return a str
    return "response here"

employee_id: int = 8

Phase 0 uses type hints on all function signatures. They cost nothing and make the code significantly more readable.

3. Functions

Functions are defined with def. Parameters are listed inside parentheses. The return value follows return. A function with no return statement returns None.

def greet(name: str) -> str:
    """Return a greeting for the given name.

    Args:
        name: The person's full name.

    Returns:
        A greeting string.
    """
    return f"Hello, {name}. How can I help you today?"


def log_ticket(ticket_id: str, category: str, priority: str = "P3") -> None:
    """Log a ticket. priority defaults to P3 if not provided."""
    print(f"Ticket {ticket_id}: {category} ({priority})")


# Call them
greeting = greet("Tom Fischer")           # → "Hello, Tom Fischer. How can I..."
log_ticket("TKT-0042", "network")         # priority defaults to "P3"
log_ticket("TKT-0043", "hardware", priority="P1")

The triple-quoted string immediately after def is a docstring. It documents what the function does, what its parameters are, and what it returns. Python tooling (IDEs, help(), documentation generators) reads docstrings automatically. Write one for every non-trivial function.

One important rule about default parameters: never use a mutable value (like a list or dict) as a default. Use None and create the mutable object inside the function instead.

# WRONG — the list is shared across all calls
def add_ticket(ticket, history=[]):
    history.append(ticket)
    return history

# CORRECT — create a new list each call when none provided
def add_ticket(ticket, history=None):
    if history is None:
        history = []
    history.append(ticket)
    return history

Input and Output

print() writes to stdout. input() reads a line from stdin and returns it as a string.

# print — handles any number of arguments
print("Hello, Tom")
print("Tickets processed:", 42)
print(f"Rate: {0.73:.1%}")     # → "Rate: 73.0%"

# input — always returns str, even if the user types a number
user_text = input("You: ")        # waits for enter key
number_text = input("Count: ")    # "15" comes back as the string "15"
count = int(number_text)          # explicit conversion required

4. Control Flow

Lists

A list is an ordered, mutable sequence. Square brackets, comma-separated values.

keywords = ["password", "vpn", "software", "hardware", "access"]

# Access by index (0-based)
first = keywords[0]    # → "password"
last = keywords[-1]    # → "access" (negative index counts from end)

# Append and remove
keywords.append("network")
keywords.remove("access")

# Length
count = len(keywords)   # → 5

# Check membership
if "vpn" in keywords:
    print("VPN is a recognised keyword")

# Iterate
for keyword in keywords:
    print(keyword)

if / elif / else

Python uses indentation to delimit blocks. There are no braces. Use four spaces per level, never tabs.

priority = "P2"

if priority == "P1":
    print("Critical — respond immediately")
elif priority == "P2":
    print("High — respond within 4 hours")
elif priority == "P3":
    print("Medium — respond within 24 hours")
else:
    print("Low — respond within 72 hours")

If you are coming from a language that uses braces for blocks, the indentation rule will feel odd for the first few hours. Python enforces it at the language level — a misindented line is a syntax error. The payoff is that all Python code looks consistent regardless of who wrote it. Configure your editor to insert 4 spaces when you press Tab — most editors do this by default for Python files.

for and while Loops

for iterates over any sequence — lists, strings, dictionary keys. while runs until a condition becomes false.

# for loop — iterates each item
categories = ["hardware", "network", "software_access"]
for category in categories:
    print(f"Processing category: {category}")

# for loop with index — use enumerate()
for i, category in enumerate(categories):
    print(f"{i}: {category}")

# while loop — runs until condition is False
count = 0
while count < 3:
    print(f"Loop iteration {count}")
    count += 1

# while True — infinite loop, exit with break
while True:
    user_input = input("You: ")
    if user_input.lower() == "quit":
        break
    print(f"You typed: {user_input}")

5. Environment Setup

Installing Python

Mac:

# Check if Python is already installed
python3 --version

# If not installed, use Homebrew
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
brew install python3

# Verify
python3 --version  # should show 3.10 or higher

Windows:

Go to python.org/downloads
Download Python 3.11 or higher
Run the installer — CHECK "Add Python to PATH"
Open Command Prompt and verify:

python --version  # should show 3.11 or higher

Setting Up Your Editor

Recommended: VS Code (free, works on Mac and Windows)

Download from code.visualstudio.com
Install the Python extension (search "Python" in Extensions)
Open a folder for your course work

Running Your First Script

Create a file called hello.py:

print("Python is working")
name = input("What is your name? ")
print(f"Hello, {name}!")

Mac: python3 hello.py

Windows: python hello.py

Installing Packages

Throughout this course you will install packages with pip:

# Mac
pip3 install package-name

# Windows
pip install package-name

If you see "command not found" on Mac, use python3 and pip3. On Windows use python and pip. Both work the same way.

6. Common Mistakes

Common Mistakes

Mistake	What Happens	Correct Approach
`if variable == None`	Works but is non-idiomatic; linters warn	Use `is None` or `is not None`
Mutable default parameter `def f(items=[])`	All callers share the same list	Use `None` as default, create mutable inside function
Forgetting `.lower()` before keyword comparison	"VPN" and "vpn" treated as different strings	Always normalise case before comparing
Using `input()` return value as a number directly	`TypeError` when doing arithmetic	Convert explicitly with `int()` or `float()`
No `if __name__ == "__main__"` guard	Top-level code runs when the file is imported	Wrap executable code in the `__main__` block
Single-character variable names (`m`, `k`, `r`)	Unreadable a week later	Use descriptive names: `message`, `keyword`, `response`
Mixing tabs and spaces	`IndentationError` at runtime	Configure your editor to insert 4 spaces, never tabs
`print` without parentheses	`SyntaxError` in Python 3 — `print` is a function, not a statement	Always use `print("text")` with parentheses

7. Practice Challenge 💪

Build a command-line expense tracker with three commands: add (expense name + amount), list (show all expenses), total (sum all amounts). Save expenses to a list in memory. No files, no classes — just functions and a loop.

What to build: A script with three commands:

add <amount> <description> — add a new expense
list — print all expenses
total — print the sum of all amounts

Requirements:

All data lives in a Python list of dicts: {"amount": 49.99, "description": "USB hub"}
add must convert the amount string to a float
list prints each expense with amount formatted to 2 decimal places ($49.99)
total prints the sum of all amounts
Unknown command prints: "Unknown command. Try: add, list, total"
Type quit to exit

Starter structure:

# expense_tracker.py — Practice Challenge, Lesson 1
expenses = []  # list of {"amount": float, "description": str}

print("IT Expense Tracker (type 'quit' to exit)")
print("Commands: add <amount> <description>, list, total")

while True:
    user_input = input("> ").strip()
    if user_input.lower() == "quit":
        break

    parts = user_input.split(maxsplit=2)
    command = parts[0] if parts else ""

    if command == "add":
        # TODO: parse amount and description, append to expenses
        pass
    elif command == "list":
        # TODO: print each expense with index and formatted amount
        pass
    elif command == "total":
        # TODO: sum all amounts and print
        pass
    else:
        print("Unknown command. Try: add, list, total")

Expected session:

IT Expense Tracker (type 'quit' to exit)
Commands: add <amount> <description>, list, total
> add 49.99 USB hub
Added: USB hub ($49.99)
> add 129.00 Mechanical keyboard
Added: Mechanical keyboard ($129.00)
> list
1. USB hub — $49.99
2. Mechanical keyboard — $129.00
> total
Total: $178.99
> quit

In Lesson 2, you refactor this into an ExpenseTracker class. In Lesson 3, you add save/load from expenses.json.

🎓

Take the quiz to test your understanding →

Create free account →

No credit card · Instant access to Phase 0

What's Next

The 20-line script works. But every time Tom wants to add a keyword or change a response, he edits the script directly — a fragile process that breaks easily when keywords overlap or response text grows. The right structure for a growing program is a class: a way to bundle related data and behavior together. Lesson 2 covers Python's data structures in depth and shows how to refactor helpdesk_v0.py into a proper HelpDeskBot class with an __init__, a process_ticket() method, and usage stats. That class structure is exactly what Phase 1 builds on when it replaces keyword matching with machine learning.

Additional Resources

Python Tutorial for Beginners (Full Course)

Video by Programming with Mosh · YouTube

🔗Useful Links

🟢 DocBot

DocBot: Your First Python Script

See how this applies →

DocBot starts even simpler than HelpDeskBot. No interactive loop — just a dictionary of answers and a function that returns the closest match by keyword.

# docbot_v0.py — Phase 0, Lesson 1
FAQ = {
    "vpn": "Connect to GlobalProtect at vpn.techcorp.com using your Okta password.",
    "password": "Reset at okta.techcorp.com. Takes 2 minutes.",
    "wifi": "Office WiFi: TechCorp-Internal. Password posted at each desk.",
    "printer": "Printer drivers at it.techcorp.com/printers. Select your floor.",
    "badge": "Lost badge? Call reception at ext. 100 immediately.",
}

def ask(question: str) -> str:
    question_lower = question.lower()
    for keyword, answer in FAQ.items():
        if keyword in question_lower:
            return answer
    return "No answer found. Try asking IT directly."

print(ask("How do I connect to VPN?"))
print(ask("I forgot my password"))

Running it:

Connect to GlobalProtect at vpn.techcorp.com using your Okta password.
Reset at okta.techcorp.com. Takes 2 minutes.

DocBot and HelpDeskBot share the same core pattern — iterate a dict, return the first keyword match. The difference is domain and interface. In Lesson 2, DocBot becomes a proper class with an ask() method and richer FAQ data. In Lesson 3, it reads its FAQ from a JSON file on disk. By Phase 4, keyword matching is replaced entirely with semantic vector search.

🔵 HelpDeskBot

HelpDeskBot: 20-Line CLI Responder

See how this applies →

The Problem HelpDeskBot Solves

Tom Fischer manages IT support for TechCorp, a 500-person technology company in Austin. Every month he handles 200 support tickets manually — password resets, VPN issues, software requests. Sarah Chen, his engineering manager, asked him to automate the simple ones.

Tom knows Java. He has heard Python is what everyone uses for AI. This is where that journey starts — not with AI, not with machine learning, just with 20 lines of Python that can respond to the five most common ticket types.

Tom's immediate problem is simple: five question types, five canned responses. Build the simplest possible thing that works. No classes, no external libraries, no configuration files. Just a dict of responses and a loop.

Here is the keyword matching flow:

HelpDeskBot v0: keyword matching loop. Each incoming message is checked against five keywords in insertion order. First match wins.

The complete script:

# helpdesk_v0.py — Phase 0, Lesson 1
# TechCorp IT HelpDesk — first 20 lines
# No classes, no AI, no external libraries. Just Python.

RESPONSES = {
    "password": "To reset your password go to okta.techcorp.com and click Forgot Password.",
    "vpn": "Download GlobalProtect from it.techcorp.com/vpn. Use your Okta credentials.",
    "software": "Submit a software request at it.techcorp.com/software-request.",
    "hardware": "For hardware issues email tom.fischer@techcorp.com with your asset tag.",
    "access": "Access requests need manager approval. Ask your manager to email IT.",
}

def get_response(message: str) -> str:
    message_lower = message.lower()
    for keyword, response in RESPONSES.items():
        if keyword in message_lower:
            return response
    return "I'm not sure how to help with that. Email tom.fischer@techcorp.com."

print("TechCorp IT HelpDesk (type 'quit' to exit)")
while True:
    user_input = input("You: ").strip()
    if user_input.lower() == "quit":
        break
    print(f"Bot: {get_response(user_input)}")

Walk through what each part does:

RESPONSES is a module-level constant — a dictionary mapping keyword strings to response strings. The uppercase name signals "do not reassign this". Python has no const keyword; naming convention carries the meaning.

get_response takes a raw message string, converts it to lowercase so "Password" and "password" both match, then iterates the RESPONSES dictionary using .items() which yields (key, value) tuples. The first keyword found in the message wins — the function returns immediately on the first match. If no keyword matches, it returns the default fallback.

The while True loop reads input indefinitely. .strip() removes leading and trailing whitespace including the newline that input() captures. The loop exits when the user types "quit". The .lower() check means "Quit", "QUIT", and "quit" all exit cleanly.

Run it:

TechCorp IT HelpDesk (type 'quit' to exit)
You: my password stopped working
Bot: To reset your password go to okta.techcorp.com and click Forgot Password.
You: how do I get on the VPN from home
Bot: Download GlobalProtect from it.techcorp.com/vpn. Use your Okta credentials.
You: I need Figma installed
Bot: Submit a software request at it.techcorp.com/software-request.
You: quit

Twenty lines. Zero external dependencies. Handles the five most common IT ticket types automatically.

What Breaks in Keyword Matching

Problem	Why It Happens	Example	Fix
Wrong keyword fires first	RESPONSES dict is ordered; first match wins	"I need access to software" matches "software" not "access"	Order keywords by priority — put more specific ones first
Case-sensitive miss	Forgot `.lower()` on message	"VPN" in message doesn't match `"vpn"` key	Always call `.lower()` before comparing
Keyword too broad	Substring match finds unintended content	`"soft"` in RESPONSES would match "soft-bricked laptop"	Use specific multi-word keywords like `"software install"`
`EOFError` crash	`input()` fails when stdin is closed (piped input)	Running `echo "quit" \| python helpdesk_v0.py` crashes	Wrap `input()` in `try/except EOFError: break`
Ugly traceback on Ctrl+C	`KeyboardInterrupt` not caught	User presses Ctrl+C to quit	Wrap the main loop in `try/except KeyboardInterrupt`
Compound queries give wrong response	Only first keyword checked	"VPN access issue" → matches "vpn" even if the real issue is "access"	Phase 1 solves this with ML classification

Design Notes

RESPONSES is a module-level dict, not hardcoded inside the function. This means all five responses are visible at the top of the file — easy to update without touching the logic. When Sarah Chen wants to change the password reset URL, she edits one line in RESPONSES. She never touches get_response.

while True with break is the standard Python idiom for interactive CLI programs. It runs until explicitly exited. The break is clearer than a boolean flag because there is no extra variable to track.

What this 20-line script handles automatically:

Password reset questions
VPN connection questions
Software installation requests
Hardware issues (routed to Tom with the right instructions)
Access requests (explained correctly on first contact)

What it cannot do yet: handle compound questions, know which employee is asking, create a ticket, or understand intent beyond exact substring matching. Those are exactly the limitations that Lessons 2 and 3 address, and that Phase 1 solves with machine learning.

By the end of this lesson you will be able to:

Declare variables and work with Python's core types (str, int, float, bool, list, dict)
Write functions with parameters, return values, and type hints
Use if/elif/else, for loops, and while loops confidently
Read input from the terminal and print formatted output
Run a Python script from the command line on Mac or Windows

1. Why Python for AI

Three reasons it matters for this course:

The ecosystem. When a new AI library ships, it ships Python first — sometimes Python only for months. If you want to use the latest tools as they arrive, you need Python.

Interactivity. Python's REPL and Jupyter notebooks let you run a single expression, see the result, modify it, run again. That tight feedback loop is how AI systems get built.

2. Variables and Types

Python is dynamically typed. You do not declare the type of a variable — you assign a value and Python infers the type at runtime.

name = "Tom Fischer"     # str
employee_id = 8          # int
confidence = 0.87        # float
is_resolved = False      # bool
assigned_to = None       # NoneType — absence of a value

The five fundamental types cover almost everything in Phase 0:

str — text. Strings use single or double quotes interchangeably. Multiline strings use triple quotes.

response = "Your password reset request has been received."
multiline = """TechCorp IT HelpDesk
Type 'quit' to exit."""

int and float — numbers. int for whole numbers, float for decimals. Arithmetic works as expected. Integer division uses //, remainder uses %.

tickets_today = 12
resolution_rate = 0.73
hours_saved = tickets_today * 0.25    # → 3.0
category_id = tickets_today // 5      # → 2 (integer division)
remainder = tickets_today % 5         # → 2

bool — True or False. Python capitalises booleans: True and False, not true and false. Comparison operators return booleans: ==, !=, <, >, <=, >=.

is_urgent = True
ticket_count = 15
needs_escalation = ticket_count > 10   # → True
is_empty = ticket_count == 0           # → False

None — the absence of a value. Use None where you would use null in Java/JavaScript or null in C#. Check for None with is None, not == None.

assigned_employee = None
if assigned_employee is None:
    assigned_employee = "tom.fischer@techcorp.com"

String Formatting: f-strings

Python has several ways to format strings. Use exactly one: f-strings. They are readable, fast, and work in every context you will encounter.

name = "Tom Fischer"
ticket_id = "TKT-0042"
status = "resolved"

# f-string: prefix the string with f, wrap expressions in {}
message = f"Hello {name}, your ticket {ticket_id} is {status}."
# → "Hello Tom Fischer, your ticket TKT-0042 is resolved."

# Expressions work inside {}
count = 7
summary = f"Processed {count} tickets. {count * 0.25:.1f} hours saved."
# → "Processed 7 tickets. 1.8 hours saved."

Type Hints

def get_response(message: str) -> str:
    # message is expected to be a str
    # this function is expected to return a str
    return "response here"

employee_id: int = 8

Phase 0 uses type hints on all function signatures. They cost nothing and make the code significantly more readable.

3. Functions

Functions are defined with def. Parameters are listed inside parentheses. The return value follows return. A function with no return statement returns None.

def greet(name: str) -> str:
    """Return a greeting for the given name.

    Args:
        name: The person's full name.

    Returns:
        A greeting string.
    """
    return f"Hello, {name}. How can I help you today?"


def log_ticket(ticket_id: str, category: str, priority: str = "P3") -> None:
    """Log a ticket. priority defaults to P3 if not provided."""
    print(f"Ticket {ticket_id}: {category} ({priority})")


# Call them
greeting = greet("Tom Fischer")           # → "Hello, Tom Fischer. How can I..."
log_ticket("TKT-0042", "network")         # priority defaults to "P3"
log_ticket("TKT-0043", "hardware", priority="P1")

One important rule about default parameters: never use a mutable value (like a list or dict) as a default. Use None and create the mutable object inside the function instead.

# WRONG — the list is shared across all calls
def add_ticket(ticket, history=[]):
    history.append(ticket)
    return history

# CORRECT — create a new list each call when none provided
def add_ticket(ticket, history=None):
    if history is None:
        history = []
    history.append(ticket)
    return history

Input and Output

print() writes to stdout. input() reads a line from stdin and returns it as a string.

# print — handles any number of arguments
print("Hello, Tom")
print("Tickets processed:", 42)
print(f"Rate: {0.73:.1%}")     # → "Rate: 73.0%"

# input — always returns str, even if the user types a number
user_text = input("You: ")        # waits for enter key
number_text = input("Count: ")    # "15" comes back as the string "15"
count = int(number_text)          # explicit conversion required

4. Control Flow

Lists

A list is an ordered, mutable sequence. Square brackets, comma-separated values.

keywords = ["password", "vpn", "software", "hardware", "access"]

# Access by index (0-based)
first = keywords[0]    # → "password"
last = keywords[-1]    # → "access" (negative index counts from end)

# Append and remove
keywords.append("network")
keywords.remove("access")

# Length
count = len(keywords)   # → 5

# Check membership
if "vpn" in keywords:
    print("VPN is a recognised keyword")

# Iterate
for keyword in keywords:
    print(keyword)

if / elif / else

Python uses indentation to delimit blocks. There are no braces. Use four spaces per level, never tabs.

priority = "P2"

if priority == "P1":
    print("Critical — respond immediately")
elif priority == "P2":
    print("High — respond within 4 hours")
elif priority == "P3":
    print("Medium — respond within 24 hours")
else:
    print("Low — respond within 72 hours")

for and while Loops

for iterates over any sequence — lists, strings, dictionary keys. while runs until a condition becomes false.

# for loop — iterates each item
categories = ["hardware", "network", "software_access"]
for category in categories:
    print(f"Processing category: {category}")

# for loop with index — use enumerate()
for i, category in enumerate(categories):
    print(f"{i}: {category}")

# while loop — runs until condition is False
count = 0
while count < 3:
    print(f"Loop iteration {count}")
    count += 1

# while True — infinite loop, exit with break
while True:
    user_input = input("You: ")
    if user_input.lower() == "quit":
        break
    print(f"You typed: {user_input}")

5. Environment Setup

Installing Python

Mac:

# Check if Python is already installed
python3 --version

# If not installed, use Homebrew
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
brew install python3

# Verify
python3 --version  # should show 3.10 or higher

Windows:

Go to python.org/downloads
Download Python 3.11 or higher
Run the installer — CHECK "Add Python to PATH"
Open Command Prompt and verify:

python --version  # should show 3.11 or higher

Setting Up Your Editor

Recommended: VS Code (free, works on Mac and Windows)

Download from code.visualstudio.com
Install the Python extension (search "Python" in Extensions)
Open a folder for your course work

Running Your First Script

Create a file called hello.py:

print("Python is working")
name = input("What is your name? ")
print(f"Hello, {name}!")

Mac: python3 hello.py

Windows: python hello.py

Installing Packages

Throughout this course you will install packages with pip:

# Mac
pip3 install package-name

# Windows
pip install package-name

If you see "command not found" on Mac, use python3 and pip3. On Windows use python and pip. Both work the same way.

6. Common Mistakes

Common Mistakes

Mistake	What Happens	Correct Approach
`if variable == None`	Works but is non-idiomatic; linters warn	Use `is None` or `is not None`
Mutable default parameter `def f(items=[])`	All callers share the same list	Use `None` as default, create mutable inside function
Forgetting `.lower()` before keyword comparison	"VPN" and "vpn" treated as different strings	Always normalise case before comparing
Using `input()` return value as a number directly	`TypeError` when doing arithmetic	Convert explicitly with `int()` or `float()`
No `if __name__ == "__main__"` guard	Top-level code runs when the file is imported	Wrap executable code in the `__main__` block
Single-character variable names (`m`, `k`, `r`)	Unreadable a week later	Use descriptive names: `message`, `keyword`, `response`
Mixing tabs and spaces	`IndentationError` at runtime	Configure your editor to insert 4 spaces, never tabs
`print` without parentheses	`SyntaxError` in Python 3 — `print` is a function, not a statement	Always use `print("text")` with parentheses

7. Practice Challenge 💪

What to build: A script with three commands:

add <amount> <description> — add a new expense
list — print all expenses
total — print the sum of all amounts

Requirements:

All data lives in a Python list of dicts: {"amount": 49.99, "description": "USB hub"}
add must convert the amount string to a float
list prints each expense with amount formatted to 2 decimal places ($49.99)
total prints the sum of all amounts
Unknown command prints: "Unknown command. Try: add, list, total"
Type quit to exit

Starter structure:

# expense_tracker.py — Practice Challenge, Lesson 1
expenses = []  # list of {"amount": float, "description": str}

print("IT Expense Tracker (type 'quit' to exit)")
print("Commands: add <amount> <description>, list, total")

while True:
    user_input = input("> ").strip()
    if user_input.lower() == "quit":
        break

    parts = user_input.split(maxsplit=2)
    command = parts[0] if parts else ""

    if command == "add":
        # TODO: parse amount and description, append to expenses
        pass
    elif command == "list":
        # TODO: print each expense with index and formatted amount
        pass
    elif command == "total":
        # TODO: sum all amounts and print
        pass
    else:
        print("Unknown command. Try: add, list, total")

Expected session:

IT Expense Tracker (type 'quit' to exit)
Commands: add <amount> <description>, list, total
> add 49.99 USB hub
Added: USB hub ($49.99)
> add 129.00 Mechanical keyboard
Added: Mechanical keyboard ($129.00)
> list
1. USB hub — $49.99
2. Mechanical keyboard — $129.00
> total
Total: $178.99
> quit

In Lesson 2, you refactor this into an ExpenseTracker class. In Lesson 3, you add save/load from expenses.json.

🎓

Take the quiz to test your understanding →

Create free account →

No credit card · Instant access to Phase 0

What's Next

Additional Resources

Python Tutorial for Beginners (Full Course)

Video by Programming with Mosh · YouTube

🔗Useful Links

🟢 DocBot

DocBot: Your First Python Script

See how this applies →

DocBot starts even simpler than HelpDeskBot. No interactive loop — just a dictionary of answers and a function that returns the closest match by keyword.

# docbot_v0.py — Phase 0, Lesson 1
FAQ = {
    "vpn": "Connect to GlobalProtect at vpn.techcorp.com using your Okta password.",
    "password": "Reset at okta.techcorp.com. Takes 2 minutes.",
    "wifi": "Office WiFi: TechCorp-Internal. Password posted at each desk.",
    "printer": "Printer drivers at it.techcorp.com/printers. Select your floor.",
    "badge": "Lost badge? Call reception at ext. 100 immediately.",
}

def ask(question: str) -> str:
    question_lower = question.lower()
    for keyword, answer in FAQ.items():
        if keyword in question_lower:
            return answer
    return "No answer found. Try asking IT directly."

print(ask("How do I connect to VPN?"))
print(ask("I forgot my password"))

Running it:

Connect to GlobalProtect at vpn.techcorp.com using your Okta password.
Reset at okta.techcorp.com. Takes 2 minutes.

🔵 HelpDeskBot

HelpDeskBot: 20-Line CLI Responder

See how this applies →

The Problem HelpDeskBot Solves

Here is the keyword matching flow:

HelpDeskBot v0: keyword matching loop. Each incoming message is checked against five keywords in insertion order. First match wins.

The complete script:

# helpdesk_v0.py — Phase 0, Lesson 1
# TechCorp IT HelpDesk — first 20 lines
# No classes, no AI, no external libraries. Just Python.

RESPONSES = {
    "password": "To reset your password go to okta.techcorp.com and click Forgot Password.",
    "vpn": "Download GlobalProtect from it.techcorp.com/vpn. Use your Okta credentials.",
    "software": "Submit a software request at it.techcorp.com/software-request.",
    "hardware": "For hardware issues email tom.fischer@techcorp.com with your asset tag.",
    "access": "Access requests need manager approval. Ask your manager to email IT.",
}

def get_response(message: str) -> str:
    message_lower = message.lower()
    for keyword, response in RESPONSES.items():
        if keyword in message_lower:
            return response
    return "I'm not sure how to help with that. Email tom.fischer@techcorp.com."

print("TechCorp IT HelpDesk (type 'quit' to exit)")
while True:
    user_input = input("You: ").strip()
    if user_input.lower() == "quit":
        break
    print(f"Bot: {get_response(user_input)}")

Walk through what each part does:

Run it:

TechCorp IT HelpDesk (type 'quit' to exit)
You: my password stopped working
Bot: To reset your password go to okta.techcorp.com and click Forgot Password.
You: how do I get on the VPN from home
Bot: Download GlobalProtect from it.techcorp.com/vpn. Use your Okta credentials.
You: I need Figma installed
Bot: Submit a software request at it.techcorp.com/software-request.
You: quit

Twenty lines. Zero external dependencies. Handles the five most common IT ticket types automatically.

What Breaks in Keyword Matching

Problem	Why It Happens	Example	Fix
Wrong keyword fires first	RESPONSES dict is ordered; first match wins	"I need access to software" matches "software" not "access"	Order keywords by priority — put more specific ones first
Case-sensitive miss	Forgot `.lower()` on message	"VPN" in message doesn't match `"vpn"` key	Always call `.lower()` before comparing
Keyword too broad	Substring match finds unintended content	`"soft"` in RESPONSES would match "soft-bricked laptop"	Use specific multi-word keywords like `"software install"`
`EOFError` crash	`input()` fails when stdin is closed (piped input)	Running `echo "quit" \| python helpdesk_v0.py` crashes	Wrap `input()` in `try/except EOFError: break`
Ugly traceback on Ctrl+C	`KeyboardInterrupt` not caught	User presses Ctrl+C to quit	Wrap the main loop in `try/except KeyboardInterrupt`
Compound queries give wrong response	Only first keyword checked	"VPN access issue" → matches "vpn" even if the real issue is "access"	Phase 1 solves this with ML classification

Design Notes

What this 20-line script handles automatically:

Password reset questions
VPN connection questions
Software installation requests
Hardware issues (routed to Tom with the right instructions)
Access requests (explained correctly on first contact)

1. Why Python for AI

2. Variables and Types

String Formatting: f-strings

Type Hints

3. Functions

Input and Output

4. Control Flow

Lists

if / elif / else

for and while Loops

5. Environment Setup

Installing Python

Setting Up Your Editor

Running Your First Script

Installing Packages

6. Common Mistakes

7. Practice Challenge 💪

Take the quiz to test your understanding →

What's Next

Additional Resources

The Problem HelpDeskBot Solves

What Breaks in Keyword Matching

Design Notes

Take the quiz to test your understanding →

Ready to continue?

Python Basics for AI Development

1. Why Python for AI

2. Variables and Types

String Formatting: f-strings

Type Hints

3. Functions

Input and Output

4. Control Flow

Lists

if / elif / else

for and while Loops

5. Environment Setup

Installing Python

Setting Up Your Editor

Running Your First Script

Installing Packages

6. Common Mistakes

7. Practice Challenge 💪

Take the quiz to test your understanding →

What's Next

Additional Resources

The Problem HelpDeskBot Solves

What Breaks in Keyword Matching

Design Notes

Take the quiz to test your understanding →

Ready to continue?