AI/ML Learning

A hands-on repository for learning machine learning, neural networks, LLM integrations, and RAG systems. The journey goes from Python and pandas fundamentals to production-ready RAG and multi-agent architectures with LangGraph.

Author: Java Backend Developer (Spring Boot, PostgreSQL, Kafka, gRPC) expanding into the AI/ML stack.

Tech Stack

• Language: Python 3.12
• ML/DL: scikit-learn, TensorFlow/Keras, PyTorch
• LLM: LangChain, LangGraph, OpenAI API, Anthropic API, DeepSeek API
• RAG: ChromaDB, Sentence Transformers, BM25, hybrid search
• Data: pandas, NumPy, Matplotlib, Seaborn
• Infra: Docker, Docker Compose, Jupyter Notebook

Project Structure

notebooks/
├── 01_basics/               # pandas, Seaborn, procurement audit analytics
├── 02_machine_learning/     # scikit-learn, classification (RandomForest, Iris)
├── 03_neural_networks/      # CNN (MNIST), RNN/LSTM (IMDB), GAN, VAE, Autoencoder
├── 04_llm_api/              # Claude API, DeepSeek API — calls and integration
├── 05_rag/                  # RAG: theory + implementation (ChromaDB, BM25, hybrid search)
├── 06_langchain/            # LangChain: chains, prompt templates, LCEL, memory, tools
├── 07_agents/               # AI agents: ReAct, tool calling, LangChain Agents
├── 08_langgraph/            # LangGraph: state graphs, checkpoints, multi-agent, supervisor
├── 09_rag_project/          # RAG chatbot: end-to-end project
├── 10_production_rag/       # Production RAG: chunking strategies, embeddings, vector stores
data/                        # Datasets (CSV)
docker/                      # Docker/Compose examples and guide
docs/                        # Learning roadmap

Modules

#	Module	Description
01	Basics	Pandas, Seaborn, visualization, procurement audit pipeline
02	Machine Learning	Classical ML: RandomForest, classification, metrics
03	Neural Networks	CNN, RNN/LSTM, Transformer, GAN, VAE, Autoencoder
04	LLM API	Working with APIs: Claude (Anthropic), DeepSeek
05	RAG	Retrieval Augmented Generation: vector and hybrid search
06	LangChain	Chains, prompt templates, LCEL, memory, output parsers, tools
07	Agents	AI agents: ReAct, tool calling, orchestration
08	LangGraph	State graphs, checkpoints, multi-agent, supervisor, subgraphs
09	RAG Project	End-to-end RAG chatbot
10	Production RAG	Chunking strategies, embedding models, vector stores

Getting Started

# Clone the repository
git clone https://github.com/rusliksu/ai-learning.git
cd ai-learning

# Install dependencies
pip install numpy pandas matplotlib seaborn scikit-learn tensorflow torch \
  langchain langchain-community chromadb sentence-transformers rank_bm25 \
  openai anthropic jupyter

# Copy the environment file
cp .env.example .env
# Add your API keys to .env

# Launch Jupyter
jupyter notebook

Status

Active learning project. Modules 01--08 are complete, 09--10 are in progress.