Frameworks and Patterns for AI Engineers
Frameworks and patterns describes the libraries and design patterns used to build agent systems, including orchestration frameworks, multi-agent layouts such as orchestrator and workers, and agents that operate a browser or computer. It helps an AI engineer decide when a framework saves effort and when plain code with a model SDK is clearer.
Stage: Agents and Tool Use · Part of the AI Engineer roadmap · Last reviewed 19 November 2026
What to learn in Frameworks and Patterns
- Compare a framework-based agent with the same agent written using only a model SDK
- Design an orchestrator that delegates subtasks to specialised worker agents
- Pass state between agents with explicit, inspectable structures
- Recognise when a multi-agent design adds cost without improving results
- Sandbox a browser-use agent and restrict which sites it can reach
Frameworks and Patterns interview questions
What does the orchestrator-worker pattern look like and when does it pay off?
A lead agent plans and splits a task, workers handle subtasks with their own context, and the lead merges the results. It helps when subtasks are parallel and independent, such as researching several sources, but it costs more tokens and needs careful merging, so simple tasks do not need it.
What extra risks does a browser or computer-use agent bring?
It acts on live pages, so prompt injection hidden in page content can redirect it, and mistakes can click buttons with real consequences. Run it in a sandbox or throwaway profile, allowlist domains, require approval for purchases or sends, and log screenshots for review.
Where Frameworks and Patterns fits in the AI Engineer roadmap
Frameworks and Patterns belongs to the Agents and Tool Use stage (intermediate level), which takes about 84 hours. Software that acts.
Other units in this stage
Back to the full AI Engineer roadmap