AI Agent Engineering: Practical Use and Design of AI Agents
Course Objective:
By the end of this course, students will be able to:
Understand the full architecture of modern AI agents
Use tools like Windsurf, Cursor, and MCP servers to configure real-world agent
workflows
Debug, evaluate, and improve agent performance
Build and deploy a fully functional AI Agent connected to external systems
Think like a systems designer: choose the right tools, prompts, and infrastructure to solve
real-world tasks
Course Structure:
Duration: 8 weeks (+ 2 half-weeks for mid-course deep dives)
Format: Weekly hands-on instruction, mini-projects, discussions
Capstone: Design and build a complete AI Agent tailored to a workflow or task of the
student’s choosing
Full Week-by-Week Breakdown
Week 0: Introduction to AI Agents in Practice
Topics:
What is an AI Agent? How is it different from chatbots or assistants?
Key components: LLM, memory, tools, logic, orchestration
Real-world examples (Devin, Monica, internal enterprise agents)
Mini Project:
🔍 Try a pre-built agent (e.g., OpenAI GPT, AgentGPT, Devin demo) to automate a task (write,
search, plan, etc.) and reflect on behavior.
Reading Materials:
1. A Visual Guide to LLM Agents by Maarten Grootendorst (Mar 2025)
https://newsletter.maartengrootendorst.com/p/a-visual-guide-to-llm-agents
2. Introduction to AI Agents: Architecture, Tools, and Implementation (Medium, Jan
2025)
https://medium.com/@aleixlopez/introduction-to-ai-agents-62a790d0bc22
3. Popular AI Agents for Devs: ChatDev, SWE-Agent & Devin (Scalable Path, Nov
2024)
https://www.scalablepath.com/machine-learning/ai-agents-chatdev-swe-agent-devin
4. Introducing Devin, the first AI software engineer (Cognition, Mar 2024)
https://www.cognition.ai/blog/introducing-devin
5. 36 Real-World Examples of AI Agents (Botpress, Jul 2024)
https://botpress.com/blog/real-world-applications-of-ai-agents
6. LLM Agent Orchestration: A Step by Step Guide (IBM, Mar 2025)
https://www.ibm.com/think/tutorials/LLM-agent-orchestration
7. Understanding LLM Agent Architectures (DataStax, Mar 2025)
https://www.datastax.com/guides/understanding-llm-agent-architectures
Week 1: Anatomy of an AI Agent
Topics:
Key architectural components:
o System prompt
o Memory/state
o Reasoning & logic
o Access to external tools/systems
o Orchestration (event loops, messaging)
Overview of agent frameworks (CrewAI, LangGraph, AutoGen, Cerebrum)
Mini Project: Dissect an open-source agent repo and sketch its architecture
Reading Materials
1. Introduction to AI Agents (Medium, Jan 2025)
https://medium.com/@aleixlopez/introduction-to-ai-agents-62a790d0bc22
2. Agent Components – Prompt Engineering Guide (Jan 2025)
https://www.promptingguide.ai/agents/components
3. What Are the Components of AI Agents? (IBM, Mar 2025)
https://www.ibm.com/think/topics/components-of-ai-agents
4. Which AI Agent Framework Should I Use? (Medium, Nov 2024)
https://medium.com/@aydinKerem/which-ai-agent-framework-i-should-use-crewai-
langgraph-majestic-one-and-pure-code-e16a6e4d9252
5. Let's Compare AutoGen, CrewAI, LangGraph, and OpenAI Swarm
(GettingStarted.ai, Oct 2024)
https://www.gettingstarted.ai/best-multi-agent-ai-framework/
6. Top 7 Frameworks for Building AI Agents in 2025 (Analytics Vidhya, Jul 2024)
https://www.analyticsvidhya.com/blog/2024/07/ai-agent-frameworks/
7. AI Agent prompt (with Manus): https://x.com/jianxliao/status/1898940029818540468
Week 1.5: GitHub & Using Open-Source Agent Frameworks
Topics:
Cloning/forking projects, running them locally
Installing dependencies, setting up .env files
Navigating repos and modifying logic
Mini Project: Fork and run an open agent repo. Customize a small piece (e.g. system prompt or
output format)
Reading materials
1. Cloning a Repository – GitHub Docs (2025)
https://docs.github.com/en/repositories/creating-and-managing-repositories/cloning-a-
repository
2. Forking and Cloning from GitHub – Steve Gordon (2025)
https://www.stevejgordon.co.uk/forking-cloning-github
3. Fork, Clone, Push, Pull: A Basic Guide to Using Git & GitHub – Medium (2025)
https://medium.com/nerd-for-tech/fork-clone-push-pull-a-basic-guide-to-using-git-github-
63ad4f5975a3
4. A Quick Review of The Most Popular AI Agent Frameworks – Medium (June 2024)
https://medium.com/@ceo_44783/a-quick-review-of-the-most-popular-ai-agent-
frameworks-june-2024-ce53c0ef809a
5. Top 5 Free AI Agent Frameworks – Botpress (December 2024)
https://botpress.com/blog/ai-agent-frameworks
6. Awesome AI Agents – Curated GitHub Repo List
https://github.com/e2b-dev/awesome-ai-agents
7. Open-Source AI Agents: How to Use Them and Best Examples – Medium (Sept
2024)
https://medium.com/@springs_apps/open-source-ai-agents-how-to-use-them-and-best-
examples-e19560280df1
8. How to Clone a GitHub Repository (YouTube Tutorial)
https://www.youtube.com/watch?v=bKuE-afbRLU
Week 2: System Prompts & Personality
Topics:
Crafting precise prompts for different use cases
Role definition, behavior shaping
Techniques: ReAct, Chain-of-Thought, Tree-of-Thought, Self-Reflection
Avoiding prompt injection
Mini Project: Write an agent system prompt that impersonates a niche role (e.g., lawyer, poet,
ruthless editor) and test responses
Reading materials
1. ReAct – Prompt Engineering Guide
https://www.promptingguide.ai/techniques/react
2. Advanced Prompt Engineering Techniques (Mercity, Feb 2025)
https://www.mercity.ai/blog-post/advanced-prompt-engineering-techniques
3. Reflexion – Prompt Engineering Guide
https://www.promptingguide.ai/techniques/reflexion
4. ReAct, Chain of Thoughts and Trees of Thoughts Explained (Medium)
https://medium.com/data-science-in-your-pocket/react-chain-of-thoughts-and-
trees-of-thoughts-explained-with-example-b9ac88621f2c
5. 5 Advanced Prompting Techniques to Improve LLM Apps (Athina, Jul 2024)
https://blog.athina.ai/5-advanced-prompting-techniques-to-improve-your-llm-app-s-
responses
Week 3: Tooling with Model-Context-Protocol (MCP)
Topics:
What is MCP (Model-Context-Protocol)?
Why it replaces function calling
Overview: model.json, context.json, protocol.json
Declarative access to data, APIs, and systems
Mini Project: Build a basic MCP config for a to-do list app (e.g., read, add, delete tasks).
Simulate how the agent would reason with it
Reading materials
1. Introducing the Model Context Protocol (Anthropic, Dec 2024)
https://www.anthropic.com/news/model-context-protocol
2. Integrating Azure AI Agents with MCP (Microsoft, Mar 2025)
https://devblogs.microsoft.com/foundry/integrating-azure-ai-agents-mcp/
3. The Model Context Protocol: Ultimate Guide (Medium)
https://medium.com/data-and-beyond/the-model-context-protocol-mcp-the-ultimate-
guide-c40539e2a8e7
4. Handle Remote Tool Calling with MCP (IBM, Jan 2025)
https://developer.ibm.com/tutorials/awb-handle-remote-tool-calling-model-context-
protocol/
5. Unlocking AI with MCP (Fleek, Mar 2025)
https://fleek.xyz/blog/learn/model-context-protocol-mcp-ai/
6. Video: understanding MCP https://x.com/aaditsh/status/1901124539553071211?
t=cphqaZ8W-kch4F1O9tkWZw&s=08
Week 4: Building an MCP-Based Agent
Topics:
Using MCP in practice
Setting up an MCP server and connecting it to a backend or data layer
Structured tool use with LLMs (via MCP schema)
Access control, data formatting, and endpoint design
Mini Project: Connect an agent to a real or dummy app via MCP (e.g., a note manager or
newsletter planner)
Week 4.5: Agent Logic & Reasoning Strategies
Topics:
What is "Agent Logic" and why it matters
Reasoning frameworks: ReAct, Planning, Self-Ask, scratchpad
Embedding logic in system prompts, memory, and orchestration
Native agent logic in newer LLMs (e.g., OpenAI o1)
Mini Project: Create a prompt scaffold that helps your agent reason step-by-step and execute a
multi-step task via MCP
Reading materials
1. Building an AI Agent with MCP and LangChain Adapters (Dev.to, Mar 2025)
https://dev.to/sreeni5018/building-an-ai-agent-with-mcp-model-context-
protocolanthropics-and-langchain-adapters-25dc
2. Model Context Protocol (MCP): The Key to Agentic AI – Video
https://www.youtube.com/watch?v=VChRPFUzJGA
Week 5: Debugging Agents & Handling Failure
Topics:
Common failure patterns: loops, hallucinations, broken tools
Debugging techniques: logs, retry logic, tool validation
Human-in-the-loop vs autonomous behavior
Agent guardrails
Mini Project: Take a buggy agent log and fix behavior by modifying reasoning, prompt structure,
or MCP actions
Reading materials
1. Building an AI Agent to Debug Deployment Errors (Ploomber, Sep 2024)
https://ploomber.io/blog/ai-debugger/
2. Using AI Agents for Notebook Debugging (JetBrains, Mar 2025)
https://blog.jetbrains.com/ai/2025/03/using-ai-agents-for-notebook-debugging
3. Debugging Your AI Agent (Cognigy, Dec 2024)
https://support.cognigy.com/hc/en-us/articles/17347113059740-Debugging-your-AI-
Agent
4. Reduce Debugging Time and Frustration With AI Agents (Zencoder, Oct 2024)
https://zencoder.ai/blog/ai-agent-reduce-debugging-time
5. AI-Powered Tools for Debugging (Medium, Nov 2024)
https://medium.com/@chirag.dave/ai-powered-tools-for-debugging-and-testing-in-
software-development-a-complete-guide-c90acda2b59b
Week 6: Evaluating & Improving Agent Output
Topics:
Evaluation methods (manual + automated)
Prompt iteration loops
Structured feedback and logging
Test cases and quality checks
Mini Project: Run a repeatable task (e.g., summarizing articles) and use eval techniques to
refine outputs
Reading materials
1. AI Agent Evaluation: Complete Overview (SuperAnnotate, Mar 2025)
https://www.superannotate.com/blog/ai-agent-evaluation
2. Evaluating AI Agents – Short Course (DeepLearning.AI, Mar 2025)
https://www.deeplearning.ai/short-courses/evaluating-ai-agents/
3. CLASSic Framework for Enterprise Agent Evaluation (Aisera, Sep 2024)
https://aisera.com/blog/ai-agent-evaluation/
4. Google’s Agent Evaluation in Vertex AI (Feb 2025)
https://cloud.google.com/blog/products/ai-machine-learning/introducing-agent-
evaluation-in-vertex-ai-gen-ai-evaluation-service
5. Survey on Evaluation of LLM-based Agents (arXiv, Mar 2025)
https://arxiv.org/abs/2503.16416
Week 7: Advanced Architectures with MCP
Topics:
Multi-agent coordination (planner/executor)
Context sharing and message passing
LangGraph, Cerebrum for stateful orchestration
Persistent agent sessions with memory
Mini Project: Build a two-agent system with defined roles (e.g., researcher + formatter) and
shared MCP context
Reading materials
1. Multi-Agent System Architecture (Smythos, Nov 2024)
https://smythos.com/ai-agents/multi-agent-systems/multi-agent-system-
architecture/
2. How Do AI Agents Handle Multi-Agent Coordination? (Milvus, Jan 2025)
https://milvus.io/ai-quick-reference/how-do-ai-agents-handle-multiagent-
coordination
3. Architectures for AI Agents (Dragonscale, Jun 2024)
https://blog.dragonscale.ai/architectures-for-ai-agents/
4. Best 5 Frameworks for Multi-Agent AI (Stream, Nov 2024)
https://getstream.io/blog/multiagent-ai-frameworks/
5. 10 Reasons Why Multi-Agent Architectures Will Supercharge AI (InfoWeek, Oct
2024)
https://www.informationweek.com/machine-learning-ai/10-reasons-why-multi-
agent-architectures-will-supercharge-ai
Week 7.5: Practical Agent Design Thinking
Topics:
How to think like an AI systems designer
Matching agent logic to task complexity
Choosing the right LLM, tools, and memory
Memory strategies (stateless, session, vector DB)
Case Examples:
Agent for website generation
Agent for coding a game
Agent for building full-stack apps
Mini Project: Design an agent blueprint for your own personal task — outline logic, tools,
context, and actions
Reading materials
1. The Architect's Toolkit: Mastering AI Agent Design Patterns (Medium, Jun 2024)
https://medium.com/@lad.jai/the-architects-toolkit-mastering-ai-agent-design-
patterns-6ee37e3c0709
2. 15 Practical AI Agent Examples to Scale Your Business (n8n, Mar 2025)
https://blog.n8n.io/ai-agents-examples/
3. 9 Real-World AI Agent Examples (Lindy.ai, Feb 2025)
https://www.lindy.ai/blog/ai-agents-examples
4. 40 AI Agent Use Cases Across Industries (Writesonic, Jan 2025)
https://writesonic.com/blog/ai-agent-use-cases
5. 30+ Agentic AI Use Cases with Real-life Examples (AIMultiple, Mar 2025)
https://research.aimultiple.com/agentic-ai/
Week 8: Capstone — Build and Deploy Your Own Agent
Goal: Use everything learned to create a fully functional AI agent connected via MCP to a real
system or workflow.
Deliverables:
A working agent
GitHub repo with documentation
Short write-up explaining design choices, tools used, and future improvements
Optional video demo
Reading materials
1. Top 10 AI Agent Projects You Can Build Today (ProjectPro, Mar 2025)
https://www.projectpro.io/article/ai-agent-projects/1060
2. AI Agent Management Strategies (Parloa, Mar 2025)
https://www.parloa.com/resources/blog/ai-agent-management-strategies-for-safe-
and-effective-deployment/
3. AI Agents: A Comprehensive Guide (Bitovi, Oct 2024)
https://www.bitovi.com/blog/ai-agents-a-comprehensive-guide
4. AI Agent Best Practices & Ethics (Writesonic, Jan 2025)
https://writesonic.com/blog/ai-agents-best-practices
5. ModelOps (Wikipedia)
https://en.wikipedia.org/wiki/ModelOps
Bonus / Optional Topics
Hosting your agent (Supabase, Vercel, local server)
UI frontends for agents (Next.js, React + API)
Using LangGraph for persistent memory + retries
Managing OpenAI / Anthropic usage costs and API telemetry
Integrating vector databases (Weaviate, Chroma)
Additional reading materials:
1. 15 Agent ideas: https://x.com/gregisenberg/status/1894031502767206599
2. AI app-building advice: https://x.com/aaditsh/status/1902019691301986654?
t=S_XDmx2XyolSQhdKR7lHDQ&s=08
3. Task management system for cursor:
https://x.com/EyalToledano/status/1903352291630961144
4. building a game with AI: https://x.com/NicolasZu/status/1898567167420809633?
t=gkFXl6u3OgNnnPk5o-4pTQ&s=08 ‘
5. vibe coding a game: https://www.youtube.com/watch?v=ulJrdYXLo9I
