LIVE AI Engineering Bootcamp: Build, Deploy & Scale AI Products

Session 2 – LLMs, Embeddings & Transformer Architecture

How does an LLM work? · Transformer Architecture · Embeddings & Semantic Similarity · Key Parameters in LLMs

Output: A working understanding of embeddings and semantic similarity

Session 3 – Vector DBs & RAG

Introduction to VectorDB · Qdrant Hands-on · RAG in Pure Python · Qdrant Advanced Operations

Output: Your First RAG Pipeline

Session 4 – LangChain, Docling & Chunking Strategies

LangChain Intro · RAG in LangChain · Docling Document Parsing · Hierarchical Chunking

Output: LangChain RAG Pipeline with Smart Chunking

Session 5 – Advanced RAG Techniques & Hands-On

Vector & Vectorless RAG Architectures · Hybrid RAG · SQL RAG · Graph RAG

Output: Project 1, LegalBot: advanced RAG over legal and financial documents with hybrid search, reranking and role-based access

Session 6 – Agentic AI Foundations

What are AI Agents? · Tool Calling & ReAct Loop · Building your first Agent with LangChain · Routing with Semantic Router & JEV · Memory in Agents

Output: Your First Tool-Calling Agent

Session 7 – Agent Orchestration & LangGraph

Introduction to LangGraph · Workflows · Conditional branches and loops · A ReAct loop in LangGraph · Hands-on

Output: A ReAct Agent Rebuilt in LangGraph

Session 8 – Multi Agent Systems

Multi-agent systems · Subagents and handoffs · Multi-agent architectures · Subagent design for CI/CD triage and PR review

Output: Project 2, DevOps agent: a multi-agent system that triages CI/CD and reviews PRs on GitHub

Session 9 – Evals, Guardrails & Observability

Online versus offline evals · Evaluating RAG pipelines · Evaluating agents · Guardrails · How it all comes together

Output: Project 3: an eval harness (20+ examples, LLM-as-judge and a guardrail) on your RAG system or agent

Session 10 – Multimodal AI

Architecture of a multimodal AI model · Extract data from PDFs/images · A multimodal RAG System

Output: A multimodal RAG system that extracts structured data from documents

Session 11 – Deployment on AWS

Introduction to AWS's Agent Stack · Deploying LangChain on AgentCore · AgentCore Services

Output: Your Agent Deployed on AWS AgentCore

Session 12 – MCP

What is Model Context Protocol · MCP Architecture · Building MCP Servers · Building MCP Clients

Output: Your First MCP Server & Client

Session 13 – Context Engineering & Harness Engineering

Why context matters in complex AI systems · Context engineering fundamentals · Context propagation in agents · Harness engineering · Harness patterns for tool use, memory and recovery · Hands-on with deepagents

Output: Project 4, IT helpdesk agent: LangGraph multi-agent + MCP server + RAG, traced in LangSmith

Session 14 – Fine Tuning & Local AI

Fine-tune vs RAG: when to opt for what? · LoRA & QLoRA training · Synthetic dataset pipelines · What are Small Language Models? · Ollama

Output: A Fine-Tuned SLM Running Locally with Ollama

Session 15 – Capstone Project Discussion & Assignment Showcase

Assignment showcase and capstone scoping

Output: Your capstone scoped, and your assignments showcased

Session 16 – AI Product Thinking and Reinforcement Learning

Turning AI capabilities into valuable products · Thinking like an AI product builder · The intuition behind reinforcement learning · Reward design

Output: A product lens on the AI systems you build

Session 17 – Advanced AI Engineering & Security

Cost optimization at scale · Semantic caching · Smart model routing · OWASP Top 10 for AI · Defending your system from adversarial attacks

Output: An AI Security & Cost Optimization Playbook

Session 18 – Stakeholder Management and System Design for AI

Managing stakeholder expectations · How an AI system is different · AI system design patterns · Production caveats

Output: AI system design patterns you can defend with stakeholders

Session 19 – Voice AI

The real-time voice stack · Low-latency voice systems · Interruptions and silence handling · Multi-turn context recovery

Output: Project 5: a voice agent under 800ms, or a multimodal agent across 3 document types

Session 20 – Inference Engineering

Runtime optimisations: batching, caching, quantization, speculation · Infrastructure: routing, load balancing, autoscaling · Serving an LLM endpoint with vLLM

Output: Project 6: a small model served with vLLM behind a cost and latency router

Session 21 – Personal Branding & Mock Interview

LinkedIn & Twitter positioning · GitHub portfolio storytelling · Live Mock Interview with Dhaval Patel & Siddhant Pandey

Output: A sharper LinkedIn and GitHub portfolio, and a live mock interview

Session 22 – Final Project Showcase

Present your final project to mentors and peers

Output: Your final project, presented to mentors and peers