Course Overview

This course bridges the gap between theoretical AI foundations and practical, real-world software engineering. You will start by understanding the mathematical and structural building blocks of neural networks and transformer architectures. From there, you will learn to programmatically orchestrate Large Language Models (LLMs), build advanced retrieval systems (RAG), design autonomous agents, and securely deploy applications to the cloud.

Format: Self-paced or Cohort-based, 100% Online

Prerequisites: Intermediate Python (functions, basic OOP) and basic data handling

Target Audience: Software engineers, data scientists, and technical product managers looking to build production-grade GenAI systems

Curriculum Roadmap

The course is organized into sequential phases, ensuring you build a solid foundation before tackling complex system architectures.

Phase 1: Foundations & Core Architectures

Weeks 1

Demystify the transition from traditional Deep Learning to Generative AI.

Topics: Artificial Neural Networks, embeddings, tokenization, and the Transformer self-attention mechanism (encoder-decoder architectures).

Key Concept: Understanding how models like GPT and LLaMA process text as statistical token predictions.

Phase 2: Applied LLM Engineering & RAG

Weeks 2

Move beyond basic prompt engineering into building context-aware applications.

Topics: Advanced prompting (few-shot, Chain-of-Thought), programmatic API integration, vector databases, and Retrieval-Augmented Generation (RAG).

Key Concept: Preventing hallucinations by grounding LLMs in custom enterprise data.

Phase 3: Agentic Workflows & Multi-Modal AI

Weeks 3

Build systems that don't just chat, but actually execute tasks autonomously.

Topics: LangChain/LangGraph orchestration, tool calling, stateful memory, and autonomous AI agents.

Key Concept: Image and multimodal generation models (Diffusion models and GANs).

Phase 4: Optimization, Deployment & MLOps

Weeks 4

Learn how to package and scale your applications for real-world traffic.

Topics: Parameter-Efficient Fine-Tuning (LoRA/QLoRA), model serving with FastAPI, containerization (Docker), and cloud orchestration (AWS Bedrock / Azure OpenAI).

Key Concept: Deploying cost-optimized, low-latency applications with guardrails against security threats like prompt injection.

Essential Tool Stack

Throughout this course, you will gain hands-on experience using industry-standard tools:

Hands-On Portfolio Projects

Rather than relying purely on multi-choice quizzes, your progress is evaluated through three milestone projects that prove your skills in real-world scenarios:

Enterprise RAG Chatbot: Create a system that ingests custom PDFs/knowledge bases, processes them into a vector database, and allows users to query the documents with high semantic accuracy.

Autonomous Task-Solving Agent: Build an AI agent utilizing LangGraph that can autonomously search the web, call external APIs, write code, and self-correct errors to solve a user-specified goal.

Fine-Tuned Specialized LLM: Take an open-source model (like LLaMA-3) and fine-tune it using LoRA on a custom, domain-specific dataset (such as medical or legal text), then deploy it via FastAPI.