Course Philosophy and Roadmap

Module 1, Lecture 1.3 | Introduction to Agentic Systems

This lecture explains the design philosophy behind the course and maps out the four-phase progression from foundations through production-ready agents. The central principle is "build it first" — every major agent component (tool registry, context management, memory, retrieval, skills) is built from scratch before any framework is introduced. We walk through what each phase covers, why coding agents serve as the primary running example, what tools you'll use (Python, Anthropic API), and what you'll be able to do by the end: evaluate frameworks, debug agent architectures, and design agents for new domains.

Read the full lecture narrative

Additional Resources

• Lecture slides
• Anthropic API — Getting Started — Official documentation for the Anthropic API, the primary LLM interface used throughout this course.
• OpenAI Platform Documentation — OpenAI's API reference; demonstrates how the patterns learned in this course transfer across LLM providers.
• Building Effective Agents — Anthropic — Anthropic's guide to agent design, emphasizing simplicity and building from composable patterns rather than reaching for complex frameworks.
• LangChain Documentation — Documentation for one of the most widely adopted agent frameworks, evaluated in Phase 3 of the course.
• CrewAI Documentation — Documentation for the CrewAI multi-agent framework, referenced in the lecture's discussion of the framework landscape.
• Retrieval-Augmented Generation (Lewis et al., 2020) — The original paper introducing RAG, a core technique covered in Phase 2 of the course.