Developing Agents an Introduction delivers a practical, code-first look at modern agentic systems in one intensive day. The course covers AI systems that reason, plan, and act through structured workflows and tool use. Participants begin by distinguishing LLMs, agents, and full agentic systems, then examine levels of agency, from single-step responses to multi-step and fully autonomous planning. The opening module also covers agent memory types, from lightweight scratchpads to episodic and long-term vector memory. It also covers the execution models, looped reasoning, event-triggered, and goal-driven, that determine when an agent decides to act.
Participants then explore how agency emerges through multi-step workflows. Prompt chaining structures multi-stage reasoning, and routing patterns direct work using rules, embeddings, or model judgment. Parallelization strategies such as branching, voting, and map-reduce reasoning speed up complex tasks. Reflection loops let an agent catch and correct its own errors, and orchestration-worker patterns split a controller’s planning from a sub-agent’s execution.
The course then focuses on tool use, covering tool calling formats and how to design reliable tool interfaces. Attendees also learn how agents reach external systems such as APIs, databases, and execution environments. This module gives particular attention to the Model Context Protocol (MCP) for secure, interoperable tool integration. It also covers safety practices such as sandboxing and rate limiting. Finally, participants learn to design and evaluate multi-step planning agents using patterns such as ReACT, Reflexion, ReWOO, and planner-executor delegation. The course closes with techniques for reducing hallucination and controlling cost and latency. Hands-on examples and guided labs let learners apply these patterns to build reliable, transparent agentic systems.
Who Should Attend
ML/AI Developers, Software & Data Engineers, IT & QA Staff, Technical Managers
What Attendees Will Learn
Upon completing Developing Agents an Introduction, participants will be able to:
- Explain how agency emerges from planning, memory, and structured workflows
- Build grounded practical agents
- Design and implement agents with multi-step workflow patterns
- Integrate and manage tools within agent systems by using tool calling, structured I/O schemas, and MCP
Prerequisites
Participants should have basic skills in Python, experience with machine learning and an understanding of LLMs.