AI Development training
Learn to integrate AI into your applications. From the fundamentals of large language models to building complete AI-powered applications and autonomous agents. Hands-on training where you build real projects.
Build the future with AI
Artificial intelligence is changing the way software is built and used. Organizations that effectively integrate AI into their applications and processes have an enormous competitive advantage. But the technology is evolving so fast that it's hard to keep up. Which models should you use? How do you integrate AI into your existing stack? And how do you build reliable, scalable AI applications that run in production?
The AI Development training from Breathbase answers these questions. In this intensive, hands-on training you learn everything you need to integrate AI into your software projects. We start with the fundamentals, so you understand how LLMs work and what their capabilities and limitations are. Then we dive into practice: you build with AI APIs, develop full stack AI applications and create autonomous AI agents that can perform complex tasks.
What makes this training special is that the trainer works with this technology daily. Breathbase builds AI-powered applications for clients and uses AI tools in every aspect of the development process. The examples, patterns and best practices you learn come directly from practice. You learn not only what is theoretically possible, but what actually works in production environments.
Training modules
From AI fundamentals to production-ready agents
- How large language models (LLMs) work: tokens, context windows, temperature
- Prompt engineering: system prompts, few-shot learning, chain-of-thought
- AI architecture overview: embeddings, vector databases, RAG patterns
- Model comparison: GPT-4, Claude, gemini, open-source models
- Limitations and risks: hallucinations, bias, security considerations
- Ethics and responsible AI use in organizations
- OpenAI API: authentication, chat completions, function calling
- Anthropic API: Claude integration, tool use, system prompts
- Streaming responses: real-time output in your application
- Structured output: JSON mode, schema validation, reliable parsing
- Integration patterns: retry logic, rate limiting, caching strategies
- Cost optimization: model selection, token management, batch processing
- Next.js + AI SDK: server actions, streaming UI, AI components
- Database design for AI: conversation storage, vector search, embeddings
- Authentication and authorization: managing API keys, user-level access
- UI/UX for AI: chat interfaces, loading states, error handling
- Deployment: Vercel, docker, environment variables, monitoring
- Testing: AI output validation, snapshot tests, integration tests
- Agent frameworks: LangChain, crewAI, vercel AI SDK agents
- Tool use: defining functions that AI can call
- Multi-step workflows: planning, execution and evaluation
- Workflow automation: triggers, conditions and AI-driven decisions
- Human-in-the-loop: approval mechanisms and oversight
- Production architecture: logging, observability and error handling
Building real projects
Practical exercises you can reuse directly after the training
Building an AI chatbot
You build a complete chatbot with streaming responses, context management and a professional UI. From API integration to deployment, you go through the complete development process and understand every step.
Implementing a RAG system
Learn to build a retrieval-Augmented generation system that allows your AI to retrieve relevant information from your own documents and databases. Essential for building AI that gives reliable, fact-based answers.
Developing an AI agent
You build an autonomous AI agent that can use tools, make decisions and execute multi-step tasks. From defining tools to implementing human-in-the-loop approval, you learn how to build reliable agents.
Who is this training for?
Suitable for technical professionals at various levels
Practical information
Flexible format, maximum results
Location
In-company or online via teams/Zoom
Duration
1 to 3 days, modularly composable
Group size
Maximum 8 participants
Materials
Code examples, templates and documentation
Requirements
Laptop, node.js, VS Code or Cursor
Aftercare
Two weeks support via email
Frequently asked questions
Related pages
View training
Explore our professional training programs in Power Platform, Dynamics 365 and AI development.