
Explore AI genetic patterns—reflection, tool use, planning, and multi-agent patterns—with hands-on sessions and practical use cases for developers, AI engineers, open minded learners, machine learning engineers, and data scientists.
Prepare by knowing a programming language—Python preferred—while grasping ai basics, machine learning, large language models, and ai agents; this course blends theory with hands-on practice.
Set up your development environment by installing Python and a code editor, obtain an OpenAI account with an API key, and use llama locally (open source) to avoid API costs.
Autogen, an open source framework for building AI agents and workflows, enables multi-agent communication, orchestration, and human-in-the-loop, with convertible agents leveraging large language models and a function and tool executor.
Discover agentic design patterns as proven blueprints for reusable, maintainable, and scalable software, solving complex object creation and system integration across databases and user interfaces, with shared vocabulary for teams.
Explore AI agentic design patterns for a large language model, revealing a workflow of reflection, planning, tool use, and multi-agent collaboration to transform user input into outputs.
The reflection pattern, a core agentic design pattern, enables iterative self-assessment of LMS outputs via self-critique and refinement, looping from initial generation through evaluation until satisfaction.
Explore the reflection design pattern workflow, where writer, critic, seo and compliance reviewers, and a meta reviewer collaborate in a nested review loop, with conversation history guiding final feedback.
Apply the reflection design pattern with Autogen to orchestrate writer and critique agents that perform compliance and seo reviews, culminating in a final draft.
Explore how to use Ollama to run open source llama models locally as a free alternative to the OpenAI API, including installation, configuration, and local querying.
Add a CEO reviewer agent to suggest SEO improvements within an agentic workflow and integrate reflection and orchestration among reviewers to deliver a grounded final article.
The reflection pattern shows a nested review workflow enabling iterative collaboration across agents, including the SEO reviewer, meta reviewer, compliance reviewer, and legal reviewer, with cost awareness for model calls.
Learn how the tool use pattern lets language models leverage external tools, such as APIs, databases, and scripts, for heavy computations and live data retrieval, enabling dynamic, modular agent workflows.
Demonstrate the tool use pattern by building a modular travel agent system with data classes for flight, hotel, and location, wired to travel tools and a local LM config.
Apply the tool use pattern by integrating planning agents, travel assistant booking agents, and flight, hotel, and location tools, replacing hard coded data with apis.
The planning pattern enables LLMs to break tasks into sequential subtasks, using the React technique of reason and act, with task analysis, subtask generation, execution in sequence, and iterative refinement.
Learn the planning pattern with the ReAct technique by building a react agent that reasons, acts, and observes using tools to plan a trip.
Plan the pattern using a react technique to map use cases for your organization, job, or projects, and reuse tool use patterns to align code with thought process.
Explore multi-agent collaboration patterns that enable task division, role assignment, agent communication, and final assembly to produce a coordinated research paper workflow.
Orchestrate multi-agent collaboration with a group chat manager and auto-generated agents to analyze papers, execute code, and generate visualizations.
Finish this course with confidence by applying AI design patterns: reflection, tool use, planning react pattern, and multi-agent collaboration to implement hands-on use cases and contribute to your projects.
Extend the projects from this course and design and implement your own agents using the tools shown here. Learn more about Autogen via Microsoft’s GitHub pages and get started online.
Transform your AI development skills by building three production-ready agent systems. Learn how Reflection, Tool Use, Planning, and Multi-Agent patterns come together to create sophisticated AI workflows using Ollama and OpenAI technologies.
Real-World Projects You'll Build
1. Research Paper Analysis Workflow
Build an intelligent system that analyzes academic papers
Implement agents that extract key findings and insights
Create automated literature review systems
Generate research summaries and annotations
Perfect for researchers, students, and academic professionals
2. AI Travel Agent System
Develop a comprehensive travel planning assistant
Create agents that handle flight and hotel recommendations
Build itinerary optimization systems
Implement budget management and scheduling
Ideal for travel applications and booking systems
3. Automated Article Writing Pipeline
Build a sophisticated content creation system
Implement research and fact-checking agents
Create style-aware writing assistants
Develop content optimization tools
Essential for content platforms and marketing teams
The Four Patterns in Action
Pattern Implementation Across Projects
Learn how each pattern drives our real-world applications:
Reflection Pattern
Research: Self-reviewing paper analysis quality
Travel: Improving recommendations based on feedback
Writing: Content quality assessment and improvement
Tool Use Pattern
Research: Integration with academic databases and citation tools
Travel: Connection to booking APIs and price comparison tools
Writing: SEO tools and content optimization platforms
Planning Pattern
Research: Structured paper analysis workflow
Travel: Intelligent itinerary creation
Writing: Content structure and article outline generation
Multi-Agent Pattern
Research: Specialized agents for different paper sections
Travel: Collaborative booking and planning agents
Writing: Research, writing, and editing agent teams
Join now to learn how to build sophisticated AI agent systems through hands-on, real-world projects. Transform theoretical patterns into practical, production-ready applications.