arrow_back All Courses
AI Development
AI Development
Stop using other people's AI products. Build your own. Connect Claude, OpenAI, and Gemini APIs. Run open-source models locally with Ollama. Implement RAG with vector search. Every module ships working PHP code on your own server. 12 practical modules, 56 lessons, and a complete multi-provider AI application deployed and owned entirely by you.
Course Curriculum
SSE (Server-Sent Events): streaming responses live
lock
Building the PHP streaming endpoint
lock
JavaScript EventSource and rendering streamed tokens
lock
Persisting conversation history in a session and database
lock
Module 2 Project: Fully functional streaming chat UI
lock
PHP and MySQL
lock
OpenAI-compatible vs Anthropic native API format
lock
Building a provider abstraction layer in PHP
lock
Adding Gemini, Grok, and Groq to the dispatcher
lock
Module 3 Project: Switch providers at runtime from a dropdown
lock
Installing Ollama on Linux and your local machine
lock
Pulling and running Llama 3.2, Mistral, and Phi-4
lock
Connecting Ollama to your PHP dispatcher
lock
Running Ollama as a systemd service
lock
Module 4 Project: Zero-cost, zero-data-leak chat with Llama
lock
Extracting text from PDF, DOCX, and CSV in PHP
lock
Sending documents to the AI API as context
lock
Summarise, extract, compare, and draft replies from docs
lock
Secure file uploads: MIME validation, size limits, storage
lock
Module 5 Project: Document Q&A interface — upload and interrogate any file
lock
What RAG is and why it matters — hallucination control
lock
Text chunking strategies: fixed, sentence, and semantic
lock
Embeddings and vector similarity search in PHP
lock
Building a knowledge base: store, retrieve, and inject context
lock
Module 6 Project: AI chatbot grounded in your own documents
lock
System prompts, personas, and output format enforcement
lock
Building a reusable prompt template engine in PHP
lock
Variable injection, validation, and preview
lock
Module 7 Project: Visual prompt builder with live output preview
lock
Multi-step generation chains: outline → draft → edit
lock
Tone and brand voice control via system prompts
lock
Generating structured output: JSON, HTML, Markdown
lock
Module 8 Project: Blog post generator from a brief
lock
DALL-E 3 via the OpenAI API from PHP
lock
Stability AI SDXL: style presets and aspect ratios
lock
AI prompt enhancement — rewrite user input for better results
lock
Module 9 Project: Image studio with style picker and gallery
lock
Encrypting API keys with AES-256-GCM in PHP
lock
Prompt injection attacks and how to prevent them
lock
Input sanitisation before and after the AI call
lock
POPIA considerations when passing user data to AI
lock
Module 10 Project: Security audit of your AI application
lock
What to log: prompt, model, tokens, latency, cost estimate
lock
Building an immutable audit log table
lock
Visualising token usage and cost over time
lock
Module 11 Project: Admin audit dashboard for all AI calls
lock
Architecture planning: modules, database schema, API structure
lock
User authentication and role-based access for AI features
lock
Integrating all modules: chat, docs, RAG, image gen
lock
Production hardening: rate limiting, error handling, logging
lock
Deployment to your VPS with database migrations
lock
Capstone showcase and peer review
lock