AI Certification Path

Most AI education is theory. Slide decks about "the future of AI" from people who haven't built anything. Peter Saddington has been running autonomous AI agents in production since 2022. Four agents. Ten websites. Daily pipelines. Real decisions. Real failures. Real learning.

This certification path is the distilled curriculum from those 4+ years. Every step has a hands-on component. You don't graduate by passing a test — you graduate by building something that works.

Step 01 — Get Set Up

Install Claude Desktop App and an AI-native IDE. Set up your development environment. No prior AI experience needed. Prerequisites only — this step gets you ready to build.

Step 02 — AI ≠ Chatbot

The most important distinction in AI: a chatbot answers questions; an agent takes action. Understand why most people use AI at 5% of its capability. Learn the problem that agents solve and why the shift from chat to action changes everything.

Step 03 — How LLMs Actually Work

Interactive deep-dive into the engine: tokenization, context windows, attention mechanisms, and prediction flow. You'll use a live tokenizer, visualize attention heatmaps, and understand why LLMs hallucinate (and how to prevent it). "Understand the engine before you drive it."

Step 04 — The 6 Agentic Capabilities

What separates an agent from a chatbot? Six capabilities:

1. Chain reasoning — Multi-step logical thinking
2. Use tools — Connect to APIs, databases, file systems
3. Self-correct — Detect and fix errors autonomously
4. Orchestrate — Coordinate multiple sub-tasks
5. Operate autonomously — Run without constant human input
6. Make decisions — Choose between options with incomplete information

Step 05 — Practice Prompting

Interactive before/after exercises showing the impact of prompt quality on output quality. Build prompting intuition through 20+ exercises across different use cases. Use the Prompt Lab to practice with ready-to-use templates organized by job function.

Step 06 — Build a Website (First Capstone)

Your first real project: use AI to build and deploy a personal website in 30 minutes. Real code. Real hosting. Real deployment. This isn't a tutorial — it's proof that you can ship with AI as your co-builder.

Step 07 — RAG Explained

Retrieval-Augmented Generation solves AI's biggest weakness: hallucination. Learn how RAG works through 12 side-by-side comparisons (with RAG vs. without). Build a chunking pipeline. Use the decision tree to know when RAG is the right tool. Peter's own AI chatbot at staas.fund runs on RAG with 2,000+ embedded documents.

Step 08 — MCP & Tool Use

How agents connect to the real world. Model Context Protocol (MCP) gives AI agents hands: email, databases, file systems, browsers, APIs. Watch tool-use flows in action, then build your own tool definition. This is where AI stops being a toy and starts being infrastructure.

Step 09 — Agent Teams & Subagents

Orchestrate multiple AI agents working in parallel. Design custom agents with specific roles and capabilities. Master 8 prompt patterns for multi-agent coordination. This is the architecture behind Peter's Council of Dogelord — 4 agents, each with a specialty, coordinating through structured communication.

Step 10 — Build 6 Agents (Final Capstone)

Hands-on capstone: build one agent for each of the six core capabilities from Step 04. By the end, you have a working multi-agent system you designed, built, and tested yourself. This is your certification artifact.

The curriculum includes 8 interactive tools that support hands-on learning at every step:

• How LLMs Work — Live tokenizer, context window visualizer, attention heatmaps
• Prompt Lab — Before/after exercises + cookbook of prompts by job function
• Whiteboard — Collaborative canvas for mapping agent flows and architectures
• Website Builder — Build and deploy a personal website in 30 minutes
• RAG Explained — 12 side-by-side comparisons, chunking demo, decision tree
• MCP & Tool Use — Watch tool-use flows, build tool definitions, explore capabilities
• Agent Teams — Direct multiple agents in parallel, design custom agents
• Agent Library — Browse 24 ready-to-use agents across 12 departments with copy-paste system prompts

Five assessment tools help you find your starting point and measure progress:

• AI Readiness Quiz — 10 questions, 2 minutes. Find your position on the 4-Stage Capability Framework
• AI ROI Calculator — Calculate hours AI can save weekly; determine what your time is worth
• What Can AI Do For You? — Select your role, get specific AI use cases with copy-paste prompts
• AI Maturity Scorecard — Rate your organization across 5 areas of AI readiness
• What Should You Automate First? — Pick your department, see highest-impact automations ranked by complexity and ROI

This isn't an academic program. There are no prerequisites beyond curiosity and a willingness to build. The curriculum works for:

• Executives who need to understand AI beyond buzzwords
• Engineers who want to build agent systems
• Entrepreneurs who want to operate at 10x scale with AI
• Teams who need a shared framework for AI adoption

Peter has delivered this curriculum to companies, conferences, and individual learners. The content is informed by 25+ years of engineering, 15+ years of Agile training, and 4+ years of daily AI agent operations.