Foundations of Deep Learning

1. Basics of neural networks

2. Activation functions

3. Loss functions & optimization

4. Backpropagation: how gradients flow

5. Overfitting, regularization, dropout

6. Batch norm, layer norm

7. Transfer learning basics

Transformers & Attention Mechanism

1. Why transformers replaced RNNs & LSTMs

2. The self-attention mechanism in detail

3. Multi-head attention: intuition & math

4. Positional encoding

5. Encoder vs Decoder: full Transformer architecture

6. Transformer variants

Large Language Models (LLMs)

1. Introduction to foundation models

2. Pre-training & fine-tuning

3. Tokenization & embeddings

4. Prompt engineering basics and advance

5. Working with locally hosted and online LLMs

6. Hugging Face transformers

Retrieval-Augmented Generation (RAG)

1. What is RAG & why it’s useful

2. Working with vector stores

3. Embedding models

4. Chunking & context windows

5. Hybrid search vs dense vector search

Agentic Frameworks

1. Concept of agents: planning, reasoning, acting

2. Autonomous vs human-in-the-loop agents

3. Tool calling & plugins

4. LangChain agents

5. LangGraph: orchestrating multi-step agent workflows

6. Memory in agents: state management

Orchestration & Advanced Pipelines

1. Chains, workflows, and graphs

2. MCP, vLLM and Ray Serve

3. Handling failures, retries, fallback models

4. Monitoring and logging agents

5. Cost and token usage management

Capstone Project

Goal:

1. Students build & deploy an end-to-end agentic AI system as a team

2. Project management with development milestones

3. Final demo