Frequently Asked Questions
Frequently Asked Questions
Common questions about the workshop, what to expect, and how to prepare.
What Python architecture does a reliable AI agent workflow need? +
A reliable Python AI agent workflow needs: a semantic blueprint generator for structured agent instructions, MCP-based agent communication with typed schemas, a RAG component for knowledge retrieval with citation tracking, a Glass-Box logging layer for observability, and safeguard components for input/output validation. This workshop builds all of these as composable Python modules.
How do I structure Python code for a multi-agent workflow? +
The workshop covers Python architecture for multi-agent workflows — how to structure agent classes with clear interfaces, how to use the MCP SDK for agent-to-agent communication, how to implement the context management layer as a reusable Python component, and how to organize the project so each agent can be tested independently before integration.
What is the best way to test a Python AI agent workflow? +
The workshop covers a testing strategy for Python AI agent workflows: unit tests for individual agent components with mocked LLM responses, integration tests for agent coordination using the MCP interface, end-to-end tests that verify the complete workflow output, and adversarial tests that check the safeguards. The Glass-Box logging layer makes each test assertion clear and debuggable.
How do I manage async operations in a Python AI agent workflow? +
The workshop covers async patterns for Python AI agent workflows — using Python's asyncio for concurrent agent invocations, managing async context managers for MCP connections, and handling async RAG retrieval without blocking the orchestration layer. Denis Rothman covers the practical async patterns that work well for production agent workflows.
Can I build a Python AI agent workflow that runs as a REST API? +
Yes. The workshop briefly covers exposing your Python AI agent workflow as a REST API endpoint — the architectural patterns for handling concurrent requests, managing agent state per request, and returning structured responses that clients can consume. This production deployment pattern is covered in module six.
What Python version and dependencies does this AI agent workflow workshop require? +
The workshop uses Python 3.10 or later. The instructor covers the specific dependencies and setup at the start of the session. All participants receive a requirements file so they can set up their environment before the workshop begins.
