Agentic AI Engineering with Python — Build Autonomous Agents Training Course

Core Concepts of Agentic AI

• Defining autonomous agents: concepts and classification
• The agent loop: the perceive, decide, act, observe cycle
• Design patterns for defining agent scope and responsibilities

Python Ecosystem and Agent SDKs

• Utilising LangChain and comparable SDKs to initialise agents
• Asynchronous programming, task queuing, and subprocess handling
• Packaging, virtual environments, and reproducible dev workflows

Integration of External Tools and APIs

• Designing tool interfaces and secure invocation methods
• Linking web APIs, databases, and internal services
• Handling credentials, secrets, and least-privilege access

Memory, State, and Context Handling

• Short-term context windows and prompt engineering methods
• Long-term memory solutions: Redis, vector stores, and retrieval augmentation
• Data consistency, caching techniques, and memory upkeep

Orchestration, Planning, and Multi-Step Processes

• Chaining actions, subagents, and task breakdown
• Comparing planning algorithms with heuristic orchestration
• Managing errors, retries, and compensatory actions

Safety, Testing, and Observability

• Threat modelling, red-teaming, and input/output sanitisation
• Unit, integration, and end-to-end testing for agents
• Logging, metrics, tracing, and alerting on agent actions

Deployment, Scaling, and Agent MLOps

• Containerisation, CI/CD pipelines, and release strategies
• Cost management, rate limiting, and resource efficiency
• Monitoring, governance, and operational playbooks

Wrap-up and Future Directions