All levels · 125 lessons

CCA-F Syllabus

125 lessons · One concept each · No experience assumed

Your progress0 / 125

Level 0

Absolute Zero

Computing fundamentals. Start here if you have zero technical background.

Week 1

What is a computer?

5 lessons

What is a computer

A machine that follows instructions — very fast, very literally, with zero judgment

What is an instruction

One single step a computer follows — not a wish, not a goal, just one action

What is running a program

Pressing 'go' — telling the computer to start following instructions now

Information stored as text, numbers, or true/false — the raw ingredients

Text vs numbers

Text is words in quotes. Numbers are values you calculate. Mixing them breaks things

Week 2

Labels, values, and JSON

5 lessons

COMPARISONL0.06

What is a label

A name tag for data — 'Name:' next to 'Danny'

What is a value

The actual data next to a label — the answer to the question

What is a key-value pair

A label and its value together: Name: Danny

What is nesting

Putting structures inside other structures — a box inside a box

Week 3

The internet and servers

5 lessons

What is the internet

Computers connected to each other to send messages back and forth

COMPARISONL0.12

What is a server

A computer that waits for questions 24/7 and sends answers back

What is a request

A message you send to a server asking it to do something

What is a response

The message a server sends back after processing your request

What is a URL / endpoint

The exact address you send a request to

Week 4

Files and folders

5 lessons

What is a list / array

An ordered collection of items — first, second, third

A named container of data saved on a computer

COMPARISONL0.18

What is a folder

A container that holds files and other folders

What is a file path

The exact address of a file — /folder/subfolder/file.txt

What is a file extension

.txt .json .md — the suffix tells you the file type

Week 5

How programs work

5 lessons

Input and output

Input = what goes in. Output = what comes out

COMPARISONL0.22

What is a function

A reusable action: inputs → does thing → outputs

What is a decision / if-then

A yes-or-no check that changes what happens next

What is repeating / iteration

Doing something over and over until a condition is met

Clicking buttons vs typing text commands

Week 6

Commands and configuration

5 lessons

What is a typed command

A text instruction typed into a terminal

What is a flag / option

A modifier that changes how a command behaves

COMPARISONL0.28

What is a setting

A saved choice that controls software behavior

Rule vs suggestion

A rule CANNOT be bypassed. A suggestion CAN be ignored

Team with a manager

One coordinator assigns work, specialists execute, results flow back

Week 7

Quality and reliability

3 lessons

Saving your progress

Recording where you are so you can resume later

Checking your work

Reviewing output to find mistakes before anyone sees them

Asking for help / escalation

Knowing when you're out of your depth and passing the problem up

Level 1

Foundation Concepts

APIs, tokens, models, and how Claude works under the hood.

Week 8

APIs and how computers talk

5 lessons

COMPARISONL1.01

A menu at a restaurant — you don't walk into the kitchen, you order from the menu

The language computers use to talk to each other — request and response, every time

The envelope format computers agree on — curly braces, keys, values, done

What is an API key

Your VIP pass — proves who you are so the server knows to let you in

What is rate limiting

The bouncer counting how many times you've come back — too many and you're cut off

Week 9

How AI models work

5 lessons

What is a model

A pattern-matching machine trained on text — not thinking, predicting

COMPARISONL1.07

What is training

Showing the model billions of examples until it learns patterns — like a chef tasting 10,000 dishes

What is a token

The smallest unit of text a model reads — not letters, not words, but chunks

What is a context window

The model's working memory — everything it can see at once, and what happens when it overflows

What is temperature

The creativity dial — 0 is robotic and predictable, 1 is creative and risky

Week 10

Prompts and responses

5 lessons

What is a prompt

The instruction you give the model — everything it knows about what you want

What is a system prompt

The invisible instruction manual the model reads before your first message

What is a completion

The model's response — it completes the pattern you started, word by word

COMPARISONL1.14

What is streaming

Getting the answer word-by-word as it's generated instead of waiting for the whole thing

What is structured output

Forcing the model to respond in a specific format — JSON, not poetry

Week 11

Claude specifically

5 lessons

Anthropic's AI model — the one this entire certification is about

How the Claude API works

Send a message, get a response — the Messages API in 10 minutes

COMPARISONL1.18

Claude model variants

Haiku, Sonnet, Opus — speed vs intelligence vs cost, pick two

The Anthropic Console

The dashboard where you manage keys, see usage, and test prompts

What is Claude Code

Claude in your terminal — reads your codebase, writes code, runs commands

Week 12

Tools and function calling

5 lessons

COMPARISONL1.21

An action Claude can take in the real world — search, calculate, look up data

What is function calling

Claude decides it needs a tool, describes what it wants, your code executes it

What is a tool schema

The instruction manual for a tool — name, description, parameters, types

Model Context Protocol — the universal standard for connecting AI to tools

The tool use loop

Message → tool call → tool result → response. The four-step dance.

Week 13

Safety and reliability basics

5 lessons

What is hallucination

When the model confidently says something that's completely wrong

What is prompt injection

When a user tricks the model into ignoring its instructions

COMPARISONL1.28

What are guardrails

The safety nets around AI — content filters, output validation, human review

What is human-in-the-loop

A human who reviews AI decisions before they become real — the safety valve

What is evaluation

Testing whether the AI actually does what you think it does — measure, don't guess

Level 2

Architecture Patterns

Prompt engineering, tool design, MCP, and agentic architecture.

Week 14

Advanced prompt engineering

5 lessons

System prompt architecture

Designing system prompts that scale — sections, XML tags, role definitions, constraints

COMPARISONL2.02

Few-shot prompting

Teaching by example — show the model 3 examples and it learns the pattern

Chain of thought

Making the model show its work — step by step reasoning before the answer

XML tags in prompts

Using XML to structure prompt sections — Claude reads tags better than headings

Prompt injection defense

Input validation, output filtering, sandwich defense, and the limits of prompt-only protection

Week 15

Tool design mastery

5 lessons

COMPARISONL2.06

Designing tool schemas

Name, description, parameters — how schema quality determines tool use quality

Tool error handling

What happens when a tool fails — retries, fallbacks, graceful degradation

Tool composition patterns

Chaining tools together — output of one becomes input of another

When to use tools vs prompts

Tools for actions, prompts for reasoning — the decision framework

Tool result formatting

How to format tool results so Claude can actually use them

Week 16

MCP deep dive

5 lessons

COMPARISONL2.11

MCP server architecture

Client, server, transport — how MCP connects models to the outside world

MCP tools vs resources

Tools do things, resources provide context — knowing which to build

MCP security patterns

Authentication, authorization, input validation — securing MCP servers

Building an MCP server

From zero to working MCP server — the minimal implementation

MCP in production

Scaling, monitoring, versioning, and what breaks when MCP servers go down

Week 17

Agentic architecture

5 lessons

What is an AI agent

A model that loops: observe, decide, act, repeat — not just chat, but do

Start, tool calls, stop conditions — designing loops that actually stop

Stop reason handling

end_turn vs tool_use vs max_tokens — the three ways Claude stops and what each means

COMPARISONL2.19

Task decomposition

Breaking complex tasks into subtasks an agent can handle one at a time

Agent guardrails

Max iterations, spending limits, approval gates — keeping agents from going rogue

Week 18

Multi-agent systems

5 lessons

The orchestrator pattern

One agent coordinates, specialists execute — the manager-worker architecture

COMPARISONL2.22

Inter-agent communication

How agents pass information — shared context, message passing, handoffs

Agent memory patterns

Short-term (context), medium-term (session), long-term (database) — the three memory layers

Parallel vs sequential agents

When to fan out (independent tasks) vs chain (dependent steps)

Agent failure modes

Infinite loops, context overflow, cascading failures, and how to handle each

Week 19

Context management

5 lessons

Token budgeting

Allocating tokens across system prompt, history, tools, and response — every token costs money

COMPARISONL2.27

Context pruning strategies

Summarization, sliding window, importance scoring — keeping context under the limit

Caching the system prompt so you don't pay for it every turn — 90% cost reduction

Conversation management

Turns, history, summarization, and what happens when a conversation gets too long

Graceful degradation

When things break, fail soft — fallback responses, cached answers, human handoff

Week 20

Production patterns

5 lessons

Claude Code configuration

CLAUDE.md, settings.json, permissions — configuring Claude Code for teams

Claude Code hooks

Pre/post hooks, custom commands, CI/CD integration — automating workflows

Monitoring AI systems

Logging, metrics, alerting — knowing when your AI is behaving badly before users do

Cost optimization

Model selection, caching, batching, prompt length — cutting your Claude bill by 80%

Security architecture

Least privilege, input sanitization, output validation — the security checklist for AI systems

Level 3

Exam Ready

Scenario-based practice, trap answers, and timed drills.

Week 21

Domain 1: Agentic Architecture (27%)

5 lessons

Agent design scenarios

3 scenarios: customer service agent, code review agent, data pipeline agent — design decisions

Enforcement pattern drill

10 scenarios: code vs prompt enforcement — which ones need locked doors?

COMPARISONL3.03

Orchestration scenarios

Single agent vs multi-agent vs no-agent — the decision tree

Safety and escalation scenarios

When does the agent stop? When does a human take over? 5 edge cases

Domain 1 trap answers

The 5 most common wrong answers in Domain 1 — and why they're wrong

Week 22

Domains 2-3: Tools & Prompts (38%)

6 lessons

Tool design scenarios

Design tool schemas for 4 real systems — payment, search, email, database

COMPARISONL3.07

MCP architecture scenarios

When to build MCP vs direct integration — 3 decision scenarios

Domain 2 trap answers

Over-engineering tools, missing error handling, wrong granularity — common mistakes

Prompt design scenarios

Design system prompts for 4 production apps — chatbot, code assistant, analyst, moderator

Claude Code config scenarios

Team setup, permissions, hooks, CLAUDE.md — 3 real-world configurations

Domain 3 trap answers

Prompt-only solutions, missing structured output, wrong model selection

Week 23

Domains 4-5: Data & Reliability (35%)

6 lessons

Data flow scenarios

Trace data through 3 AI systems — input, processing, output, storage

COMPARISONL3.13

JSON schema scenarios

Design JSON schemas for tool inputs, API responses, and structured output

Domain 4 trap answers

Type mismatches, missing validation, wrong data format — the data traps

Reliability scenarios

Context overflow, hallucination detection, cost explosion — 5 production emergencies

Evaluation design scenarios

Build eval sets for 3 AI features — what to measure, how to measure, when to fail

Domain 5 trap answers

Trusting vibes over metrics, testing happy path only, ignoring cost

Week 24

Cross-domain scenarios

5 lessons

Scenario: Customer service AI

Full architecture for a customer service agent — every domain in one system

Scenario: Code review pipeline

Claude Code + CI/CD + multi-agent review — a real production pipeline

Scenario: Medical triage AI

High-stakes AI with human-in-the-loop — where every domain matters

COMPARISONL3.21

Scenario: Data extraction pipeline

Documents in, structured data out — multi-step extraction with validation

Scenario: Content moderation

Scale, speed, accuracy, appeals — moderating millions of posts per day

Week 25

Exam strategy and final review

5 lessons

Exam format deep dive

60 questions, 4 random scenarios, time management — the tactical playbook

COMPARISONL3.24

Answer elimination strategy

How to eliminate wrong answers — the prompt vs code test, the specificity test, the scope test

The 10 exam patterns

The 10 question patterns that appear on every CCA-F exam — recognize them, answer them

Practice exam walkthrough

20 questions with detailed explanations — simulate the real exam experience

Final review and exam day

Everything you need to know in one lesson — the night-before cheat sheet

About the CCA-F exam

60 questions · 120 minutes · 720/1000 to pass · $99 (first 5,000 free via partner network) · 5 domains · 6 scenarios (4 random per exam) · No penalty for guessing

Study with us

Get a daily CCA-F concept + exam tip by text. Free. One text per day.

1 text/day. Unsubscribe anytime.