Back to Courses
DevelopmentIntermediate56 lessons9 modules
AI Agents Masterclass
A focused, practical introduction to AI agents. Build 3 working agents in 20 hours. For the comprehensive 65-hour deep dive with 4 frameworks and 4 projects, see AI Agents from Scratch.
$49per month — this curriculum and the other 13
Courses are not sold one at a time. One subscription opens this curriculum and every other one, each with its own coach. See the full terms. Card required to start the trial; the first payment lands on day 8 and renews until you cancel.
What's Included
- The full curriculum — 9 modules, 56 lessons
- An AI coach that knows this curriculum and your business
- Progress tracking across the curriculum
No videos, no written lesson pages, no quizzes, no certificate. You work through the curriculum in conversation with the coach.
7-Day Money-Back Guarantee
Not satisfied? Get a full refund within 7 days. No questions asked.
What You'll Learn
Design agent architectures using tool use, memory, and planning loops
Implement function calling and tool integration with any LLM provider
Build persistent memory systems with vector stores and conversation history
Create planning and reasoning pipelines that break complex tasks into steps
Orchestrate multi-agent systems with delegation and coordination patterns
Deploy agents to production with Docker, cloud functions, and queue systems
Set up monitoring, logging, and observability for autonomous agents
Implement safety guardrails, sandboxing, and human-in-the-loop controls
Outcomes
- Build autonomous agents with tool use, memory, and planning capabilities
- Design and deploy multi-agent systems for real business tasks
- Implement production patterns: error handling, rate limiting, monitoring
- Apply safety guardrails and evaluation frameworks to AI agents
Prerequisites
- -Python or JavaScript fundamentals
- -Basic understanding of AI APIs
Projects You'll Build
- Build a tool-using agent with persistent memory
- Create a multi-agent customer support system
- Deploy an agent with production monitoring
Course Curriculum
Module 1: Agent Fundamentals
- 1.1What makes an AI agent different from a chatbot
- 1.2Building your first agent in 30 minutes
- 1.3The agent loop: perceive, reason, plan, act, observe
- 1.4Agent architectures: reactive, deliberative, and hybrid
- 1.5Choosing the right LLM for your agent (cost vs capability)
- 1.6Setting up your agent development environment
Module 2: Tool Use & Function Calling
- 2.1Why tools are the key to useful agents
- 2.2Defining tool schemas with JSON Schema
- 2.3Implementing tool execution with error handling
- 2.4Building common tools: web search, file I/O, database queries
- 2.5Parallel and sequential tool execution patterns
- 2.6Dynamic tool selection — letting agents choose the right tool
- 2.7Testing and debugging tool integrations
Module 3: Memory & State Management
- 3.1Why agents need memory beyond the context window
- 3.2Short-term memory: sliding windows and summarization
- 3.3Long-term memory with vector databases
- 3.4Episodic memory: learning from past successes and failures
- 3.5Shared state management for multi-session agents
- 3.6Memory retrieval strategies: recency, relevance, importance
- 3.7Implementing memory with Redis, PostgreSQL, and Pinecone
Module 4: Project 1 — Tool-Using Agent with Persistent Memory
- 4.1Architecture design: tool selection, memory layer, planning loop
- 4.2Building the tool suite: web search, file ops, and database queries
- 4.3Integrating persistent memory for cross-session context
- 4.4Testing and iterating on agent behavior
- 4.5Packaging and documenting your agent
Module 5: Multi-Agent Systems
- 5.1When single agents aren't enough
- 5.2Orchestrator-worker pattern: one agent delegates to many
- 5.3Peer-to-peer agent communication protocols
- 5.4Consensus and voting patterns for agent decisions
- 5.5Shared context and message passing between agents
- 5.6Building a research team: planner, searcher, writer, reviewer
- 5.7Handling conflicts and deadlocks in multi-agent workflows
- 5.8Testing multi-agent coordination end-to-end
Module 6: Project 2 — Multi-Agent Customer Support System
- 6.1Architecture: router, FAQ, ticket, and escalation agents
- 6.2Building the knowledge base and FAQ retrieval layer
- 6.3Implementing sentiment-based routing and escalation logic
- 6.4End-to-end testing of the multi-agent support flow
- 6.5Deploying the system and monitoring conversations
Module 7: Deployment & Infrastructure
- 7.1Packaging agents for deployment with Docker
- 7.2Serverless agent deployment on AWS Lambda and Vercel
- 7.3Queue-based architectures for long-running agent tasks
- 7.4Scaling agents: concurrency, rate limits, and cost budgets
- 7.5CI/CD pipelines for agent code and prompt updates
- 7.6Infrastructure as code for agent systems
Module 8: Production Patterns & Safety
- 8.1Input validation and prompt injection defense
- 8.2Output guardrails: content filtering and format validation
- 8.3Sandboxing agent actions: limiting file system and network access
- 8.4Human-in-the-loop: approval workflows for high-risk actions
- 8.5Monitoring agent behavior: anomaly detection and alerting
- 8.6Cost management: token budgets, caching, and model fallbacks
- 8.7Evaluation frameworks: measuring agent accuracy and reliability
- 8.8Incident response: what to do when agents go wrong
Module 9: Project 3 — Production Agent with Monitoring
- 9.1Designing an agent for a real-world use case of your choice
- 9.2Implementing production-grade error handling and fallbacks
- 9.3Setting up logging, tracing, and alerting dashboards
- 9.4Your Agent Portfolio — 3 agents deployed and documented
Ready to Start Learning?
You bring a lesson from the curriculum, your coach works through it with you against your actual situation, and you keep whatever you build.
7-day money-back guarantee — no questions asked
Start where you are
Stop watching tutorials. Start building.
Pick a curriculum and your coach works through it against your own business. The starter course is free — no card — so you can see how the coaching actually feels before you decide anything.