Every AI system is still just a program underneath. We make sure that foundation is solid before touching a single model.

• Python fundamentals, data structures & OOP for AI systems
• NumPy & Pandas for data handling
• APIs, JSON & system integration
• File handling, automation scripts & modular code
• Debugging, logging & configuration workflows

Project: Reusable data-processing pipeline PythonNumPyPandas

Before you can build an agent, you need to understand what you're actually talking to, and how to talk to it well.

• Transformers, attention & tokenization
• How LLMs generate text: context windows & next-token prediction
• Prompt engineering: zero-shot, few-shot, chain-of-thought
• System prompts, role based prompting & structured output
• Function calling and structured JSON responses
• Comparing model choices: GPT, Claude, Gemini, Llama & Granite

OpenAI APIClaude APIGemini APIHugging Face

An LLM only knows what it was trained on. This module is about connecting it to your own, current data, accurately.

• Generating and comparing embeddings
• Chunking strategies & data ingestion
• Vector databases: FAISS, ChromaDB & Pinecone
• Retrieval pipelines & semantic search
• Reranking, hybrid retrieval & reducing hallucinations
• Agentic RAG: retrieval inside a reasoning loop

Project: Production-grade RAG assistant FAISSChromaDBLangChain

This is where the program turns a corner, from calling a model to building something that reasons and acts on its own.

• What makes a system "agentic" versus a simple chatbot
• Tool calling & chaining with LangChain Expression Language (LCEL)
• Extracting tool inputs from LLM outputs, then validating and executing the call
• ReAct style reasoning: think → act → observe
• Using built-in LangChain agents to analyze data, generate visualizations & run database queries
• Short-term & long-term memory for agents
• Designing and binding custom tools safely: accuracy, safety & cost tradeoffs

Project: an agent that executes tasks with real tool access LangChainLCEL

Single agents hit a ceiling fast. LangGraph is how you give a system memory, structure, and the ability to correct itself as it works.

• Workflows built as graphs: state, nodes & edges
• Conditional logic & iteration in agent workflows
• Agents that improve themselves: Reflection, Reflexion & ReAct architectures
• Structuring agent feedback and integrating external data while reasoning
• Human-in-the-loop checkpoints for higher-stakes decisions
• Agentic RAG systems that route queries and support reasoning enhanced by retrieval
• Agent orchestration, query routing & governance for reliability at scale

Project: multi-step research agent with self-correction LangGraph

One agent can only do so much. This module is about splitting work across a team of agents the way you'd delegate across a real team.

• Structuring agents, tasks & tools into crews
• Designing agent roles and delegating tasks
• Selecting and combining agentic frameworks & architectural patterns for a given problem
• Memory management across multiple collaborating agents
• Agents driven by conversation, built with AG2 (AutoGen) and BeeAI
• Agent-to-agent communication patterns
• Comparing frameworks (LangGraph vs. CrewAI vs. AG2 vs. BeeAI) for real-world fit

Project: Collaborative multi-agent system CrewAIAutoGen (AG2)BeeAI

Most companies hiring for agentic AI roles already run on Azure. This module makes sure you can ship there too.

• Azure AI Foundry: projects, model deployment & managed identity
• Implementing generative AI & agentic solutions on Azure
• Grounding agents with Azure AI Search & RAG patterns
• Responsible AI, content filtering & security on Azure
• Practice aligned to the exam: planning & managing Azure AI solutions

Azure AI FoundryAzure AI SearchAzure OpenAI Service

MCP is quickly becoming the standard way agents connect to tools and data. This module doubles as direct prep for Anthropic's own architecture exam.

• Model Context Protocol: servers, clients, tools, resources & prompts
• Connecting models, tools and data through MCP
• Claude Agent SDK & Claude Code workflows
• Structured output design and context window management
• Reliability patterns: multi-instance review and coherence across long conversations
• Mapped to Anthropic's Claude Certified Architect – Foundations exam domains

Project: MCP based workflow automation system MCPClaude APIClaude Agent SDK

A working notebook isn't a product. This is the module where your agent becomes something other people can actually rely on.

• Serving agents with FastAPI & containerizing with Docker
• Deploying to the cloud & designing for scale
• Prompt versioning, tracing & observability
• Evaluation pipelines, guardrails & hallucination checks
• Caching, batching & cost/latency optimization
• CI/CD for AI applications

Project: Deployed, monitored production agent FastAPIDockerLangSmith

• End-to-end capstone: build & ship a multi-agent product
• Mock interviews with AI engineering hiring panels
• Sample exam papers for CCA-F and AI-103
• Resume & LinkedIn profile built around your project portfolio