Bootcamp · 12 modules
AI Agent Engineering
Build production-ready AI agents from scratch — from LLM fundamentals to autonomous, multi-agent pipelines.
LLM basics-ல இருந்து RAG, tool calling, MCP, multi-agent வரைக்கும் — சொந்தமா ஒரு AI agent-ஐ உருவாக்கி deploy பண்ணலாம்.
About this program
Learn how modern AI systems are architected, evaluated and deployed. Go step by step from LLM fundamentals to autonomous, multi-agent pipelines, and finish with a deployed AI agent portfolio project.
Who it is for
- Developers who want to build with LLMs
- Backend engineers adding AI features to products
- Students who want an AI portfolio project
What you will be able to do
- Write reliable prompts and structured outputs
- Build tool-calling workflows and agents with memory
- Build advanced RAG systems and multi-agent pipelines
- Build and use MCP servers
- Deploy with FastAPI, Docker and AWS, and observe it with tracing and evals
Syllabus
Python Essentials for AI Engineers4 topics
- Fundamentals, Functions & Modules
- OOP & Virtual Environments
- APIs & Async Python
- Project Structure & Pydantic
AI & Large Language Models4 topics
- AI vs ML vs Deep Learning vs GenAI
- Tokens, Tokenization & Transformers
- Attention Mechanisms & Context Windows
- OpenAI, Claude, Gemini & Open Models
Prompt Engineering4 topics
- System & User Prompt Architecture
- Few-Shot & Zero-Shot Strategies
- Chain of Thought (CoT) & Templates
- Prompt Evaluation & Structured Output
Principles of AI Applications4 topics
- AI Apps vs AI Agents
- Context Engineering & Assembly
- JSON Mode & Schema Validation
- Building Reliable AI Systems
Tool Calling & Workflows4 topics
- Function Calling & Custom Tools
- External API & Database Tools
- Human-in-the-Loop Integrations
- Sequential, Parallel & Conditional Workflows
Memory for AI Agents4 topics
- Short-Term & Long-Term Memory
- Semantic vs Episodic Memory
- Conversation Management & Optimization
- Personalization Architectures
Retrieval-Augmented Generation4 topics
- Chunking, Embeddings & Vector DBs
- Parent Document & Multi-Vector Retrievers
- Hybrid Search, Query Transform & Re-ranking
- Semantic Caching & Hallucination Control
Building AI Agents4 topics
- Planning, Reasoning & Reflection
- Autonomous Loops & Execution
- Failure Recovery & Human Feedback
- Reliable Agent Design
Multi-Agent Systems4 topics
- Planner, Worker, Reviewer & Coordinator Roles
- Agent-to-Agent Communication
- Shared Memory & Distributed Workflows
- Collaborative Problem Solving
Model Context Protocol (MCP)4 topics
- MCP Clients & Server Architectures
- Resources, Prompts & Tools
- Building Custom MCP Servers
- Integrating MCP into Production Agents
Production AI Engineering4 topics
- FastAPI, PostgreSQL & Redis
- Docker Containerization & AWS Deployment
- Authentication & Secrets Management
- Cost Optimization & Scaling
Evaluation & Observability4 topics
- Benchmarking & LLM-as-a-Judge
- Guardrails & Prompt Testing
- Tracing, Logging & Token Analytics
- Debugging Production AI Apps
Projects
- Capstone: a production-ready AI agent platform with LLM integration, tool calling and memory, advanced RAG, multi-agent systems, MCP servers, a FastAPI backend, Docker and AWS hosting, and observability and tracing