05 - AI Agents
What Are AI Agents
AI Agents are LLM-powered systems that can perceive their environment, reason about goals, and take actions using tools - going beyond single-turn chat to autonomous, multi-step task completion.
What You'll Learn
| # | Concept | Description |
|---|---|---|
| 01 | What Are AI Agents | Definition, the four pillars, and the autonomy spectrum |
| 02 | Anatomy of an AI Agent | Deep dive into the six core components: Brain, Planning, Tools, Memory, Perception, and Governance |
| 03 | Agent Memory | The four memory types, the memory lifecycle, and how agent memory differs from workflow memory |
| 04 | AI Agent Capabilities | Perception, reasoning, planning, memory, tool use, and self-correction |
| 05 | AI Agent Use Cases | When to use AI Agents; domain-specific examples and decision framework |
| 06 | Enterprise vs Personal AI Agents | Production requirements, security, governance, and framework selection |
| 07 | Agent Evaluation Basics | Single-agent test sets, tool-call and output checks, lightweight LLM-as-judge, when to graduate to system-level eval |
Hands-On Labs
Build AI Agents using four major frameworks, progressing from simple to complex:
| Framework | Best For | Labs |
|---|---|---|
| GCP Agent Development Kit | Google Cloud-native AI Agents | Fundamentals → Simple → Complex |
| LangChain | General-purpose AI Agent chains | Fundamentals → Simple → Complex |
| LangGraph | Stateful, graph-based AI Agents | Fundamentals → Simple → Complex |
| CrewAI | Role-based multi-AI-Agent crews | Fundamentals → Simple → Complex |
See AI Agent Types for a complexity progression overview across all four frameworks.
Framework Comparison
| Dimension | GCP ADK | LangChain | LangGraph | CrewAI |
|---|---|---|---|---|
| State management | Session-based | Memory modules | Graph state | Shared crew context |
| Multi-AI-Agent | Via orchestration | Via chains | Via graph nodes | Native (Crew) |
| Cloud integration | GCP native | Cloud-agnostic | Cloud-agnostic | Cloud-agnostic |
| Best complexity | Medium–High | Low–Medium | Medium–High | Medium–High |
Recommended Learning Path
- What Are AI Agents - understand the definition and core pillars
- Anatomy of an AI Agent - understand the six core components in depth
- Agent Memory - dive deep into how agents remember and learn over time
- AI Agent Capabilities - learn what makes an AI Agent capable
- AI Agent Use Cases - understand where AI Agents add the most value
- Enterprise vs Personal AI Agents - understand production constraints
- Agent Evaluation Basics - learn to test and evaluate a single agent before scaling up
- Pick one framework and work through its Fundamentals → Simple AI Agent → Complex AI Agent labs
- Return to Interview Q&A to consolidate knowledge