Generative AI and RAG Fundamentals

Last Updated: June 8, 2026

Comprehensive coursework covering Generative AI, Retrieval-Augmented Generation (RAG) systems, and production-grade pipelines. This repository contains concept-focused notes, hands-on notebooks, and resources for building correct mental models before extending into enterprise AI applications.

Tags: generative-ai, rag, llm, nlp, vector-search, chromadb, bert, langchain, ai-engineering, machine-learning, peft, lora, qlora, fine-tuning

Learning Path

00 Docker Fundamentals ─────────── Containerization for reproducible environments
        │
01 BERT & NLP Foundations ──────── NLP mental models with transformers
        │
02 RAG & LLM Fundamentals ─────── Retrieval-augmented generation concepts
        │
03 Vector Search & Retrieval ───── Efficient semantic search with ChromaDB
        │
04 Advanced RAG & Fine-Tuning ──── LangChain orchestration, evaluation,
        │                           guardrails, LoRA/QLoRA fine-tuning
        │
05 Data Engineering ────────────── Scalable data pipelines with Kafka & Spark

Directory Structure

00-Docker-Fundamentals

Container infrastructure and deployment basics.

File	Description
Docker_Basics.pdf	Introduction to Docker, containerization concepts, and image management
Docker_Networks_and_Composition.pdf	Docker networking, multi-container orchestration, and Compose workflows

01-BERT-NLP-Foundations

Foundational NLP models and transformer architecture.

File	Description
bert_fundamentals.ipynb	Tokenization, loading BERT, extracting hidden states, NER, custom PyTorch structures
bert_flashcards.html	Visual explanation of BERT encoder blocks, MLM, and NSP

02-RAG-LLM-Fundamentals

Core concepts for prompting, generation, embeddings, and basic similarity matching.

File	Description
llm_generation_embeddings.ipynb	Embeddings via APIs (OpenAI/OpenRouter), text generation with GPT-2
rag_scratch_implementation.ipynb	Complete RAG loop with pure math (cosine similarity, dot products)
hugging_face_ecosystem_walkthrough.ipynb	Walkthrough of the Hugging Face ecosystem
Hugging_Face.pptx	Presentation on Hugging Face libraries and models
rag_01_foundations_and_architecture.pdf	Why RAG exists, core architecture, retrieval-generation flow
rag_03_query_time_retrieval.pdf	Query-time retrieval, prompt augmentation, grounded generation

03-Vector-Search-and-Retrieval

Vector search algorithms, ChromaDB implementations, and similarity-based retrieval.

File	Description
chromadb_walkthrough.ipynb	Vector search using ChromaDB
Vector_Search_Fundamentals.pptx	Vector search concepts and applications
Chunking_Hybrid_RAG_Evaluation.pdf	Chunking strategies, hybrid RAG, retrieval evaluation metrics

04-Advanced-RAG-LangChain

LangChain orchestrations, retrieval evaluations, guardrails, and parameter-efficient fine-tuning.

Notebooks:

#	File	Description
01	rag_knowledge_graph_neo4j.ipynb	Graph-based RAG using Neo4j and LangChain
02	advanced_rag_self_query.ipynb	Self-querying retrieval with metadata filtering
03	rag_evaluation.ipynb	RAG evaluation using RAGAS metrics (faithfulness, relevancy, context precision)
04	chatbot_evaluation.ipynb	Evaluating conversational AI with automated metrics
05	llm_guardrails.ipynb	Safety guardrails for LLM applications (input/output validation, content filtering)
06	lora_qlora_peft_finetuning.ipynb	LoRA, QLoRA, PEFT fine-tuning with OPT-350M on WikiText-2 — architecture diagrams, weight freezing, gradient checkpointing, adapter management

Reference Materials:

File	Description
KV_and_Graph_DB_Notes.pdf	Key-value stores, graph databases, structured data retrieval
AWS_EC2.pdf	EC2 instance types, configuration, and deployment for AI workloads
Pinecone__Vector_Database.pdf	Pinecone vector database architecture, indexing, production deployment

Note: Neo4jVector import has moved. Use:

from langchain_neo4j.vectorstores import Neo4jVector  # pip install langchain-neo4j

05-Data-Engineering-and-Streaming

Real-time data processing and complementary big data technologies.

File	Description
kafka_pyspark_streaming_walkthrough.ipynb	Real-time pipelines with Apache Kafka and PySpark Streaming (Windows setup)
Big_Data_Technologies.pdf	Big data platforms, distributed computing, data processing frameworks
Time_Series_Fundamentals.pdf	Time series analysis, forecasting, temporal data patterns
Database_Management_Fundamentals.pdf	Database concepts, ACID properties, query optimization

Learning Goal

This repository focuses on building correct mental models first, followed by lightweight hands-on practice, extending natively into production-grade AI pipelines and enterprise applications.