Start
AI Engineering
A book about how to make AI systems suitable for production.
ML and AI are becoming the infrastructure for modern products - from search and recommendations to assistants and automation. To design such systems, you need to understand not only the model, but also data, cost, latency, security and scaling. This section helps you develop an engineering perspective on AI systems and prepare you for how they impact a developer's career.
Why is this important for systems design?
AI as part of the product
Recommendations, search, assistants - increasingly, they are the ones who create value for the user.
Engineering compromises
Model quality, latency, cost of inference and availability are classic trade-offs.
Data = new infrastructure
You need to understand data pipelines, quality control, and sourcing responsibility.
Risk management
Security, privacy and error tolerance are part of the architecture, not “later”.
Features and Limitations
Possibilities
- Behavior-based personalization and recommendations.
- Intelligent search, ranking and classification.
- Automation of routine actions and decision support.
- New interface for communicating with the system via text and voice.
Restrictions
- The quality and relevance of the data directly affects the results.
- Inference can be expensive and slow without optimizations.
- Models make mistakes and require explainability and control.
- Security, privacy and data bias risks.
Section map: what's inside
AI Engineering
How to build AI products: from RAG to quality assessment and production practices.
LLM and internships
Prompting, context, agents and workflows.
ML ecosystem
Practices and tools that brought ML to the masses.
Stories and context
How we came to modern AI systems.
Why is this for an engineering career?
- Understanding the AI stack expands the range of projects and roles.
- The ability to evaluate the value of AI features helps make product decisions.
- Knowing the limitations reduces the risk of “magical expectations” from models.
- An AI-powered engineer remains competitive in the market.