AI Engineering Hub

Production-grade AI engineering patterns, examples, and best practices for building LLM-powered applications.

Overview

This repository demonstrates modern AI engineering practices including:

• Multi-Provider LLM Support - Unified interface for OpenAI, Anthropic, and other providers
• Production Patterns - Rate limiting, retry logic, structured logging, error handling
• RAG Implementation - Vector search, document chunking, context augmentation
• Agent Architectures - ReAct pattern, tool use, multi-step reasoning
• Evaluation Framework - LLM-as-judge, automated quality assessment
• Hybrid ML/LLM Pipelines - When to use traditional ML vs. LLMs

Quick Start

# Clone the repository
git clone https://github.com/MysterionRise/ai-engineering-hub.git
cd ai-engineering-hub

# Install dependencies
pip install -e ".[dev]"

# Set up your API keys
export OPENAI_API_KEY="your-key-here"
export ANTHROPIC_API_KEY="your-key-here"  # Optional

# Run an example
python examples/01-llm-fundamentals/basic_chat.py

Project Structure

ai-engineering-hub/
├── src/ai_hub/              # Core library
│   ├── core/                # Config, errors, logging, retry utilities
│   ├── providers/           # LLM provider abstractions
│   └── utils/               # Token counting, text utilities
├── examples/                # Runnable examples by category
│   ├── 01-llm-fundamentals/ # Chat, streaming, structured output
│   ├── 02-multi-provider/   # Provider comparison and fallbacks
│   ├── 03-function-calling/ # Modern tools API usage
│   ├── 04-rag-patterns/     # Retrieval-augmented generation
│   ├── 05-agents/           # ReAct and agent patterns
│   ├── 06-evaluation/       # LLM-as-judge evaluation
│   ├── 07-fine-tuning/      # Data prep for fine-tuning
│   ├── 08-vision-multimodal/# Image analysis with Vision API
│   ├── 09-traditional-ml/   # When ML beats LLMs
│   └── 10-production-patterns/ # Rate limiting, caching
├── projects/                # Complete mini-projects
│   └── semantic_search/     # Full semantic search implementation
├── tests/                   # Unit and integration tests
└── docs/                    # Documentation

Example Gallery

🎯 Fundamentals

Example	Description	Difficulty
Basic Chat	Single and multi-turn conversations	⭐
Streaming	Real-time response streaming	⭐
Structured Output	JSON mode and Pydantic validation	⭐⭐

🔄 Multi-Provider

Example	Description	Difficulty
Provider Comparison	Compare responses across providers	⭐⭐

🔧 Function Calling

Example	Description	Difficulty
Basic Tools	Modern tools API with execution loop	⭐⭐

📚 RAG Patterns

Example	Description	Difficulty
Simple RAG	Vector search with context augmentation	⭐⭐⭐

🤖 Agents

Example	Description	Difficulty
ReAct Agent	Reasoning and acting pattern	⭐⭐⭐

📊 Evaluation

Example	Description	Difficulty
LLM-as-Judge	Automated quality evaluation	⭐⭐⭐

🏭 Production

Example	Description	Difficulty
Rate Limiting	Token bucket rate limiter	⭐⭐⭐

Core Library Usage

from ai_hub import OpenAIProvider, Message

# Initialize provider
provider = OpenAIProvider(default_model="gpt-4o")

# Simple completion
response = provider.complete([
    Message.system("You are a helpful assistant."),
    Message.user("What is the capital of France?"),
])
print(response.content)

# With streaming
for chunk in provider.stream([Message.user("Tell me a story")]):
    print(chunk.content, end="", flush=True)

# Multi-provider support
from ai_hub import get_provider

openai = get_provider("openai")
anthropic = get_provider("anthropic")  # Requires anthropic package

Architecture Principles

1. Provider Abstraction - Write once, run on any LLM
2. Production-Ready - Retry logic, rate limiting, structured logging
3. Type Safety - Full type hints with Pydantic validation
4. Testability - Mockable interfaces, comprehensive test coverage
5. Observability - Structured JSON logging with metrics

Configuration

Configuration via environment variables (or .env file):

# Required
OPENAI_API_KEY=sk-...

# Optional
ANTHROPIC_API_KEY=sk-ant-...
AI_HUB_DEFAULT_MODEL=gpt-4o
AI_HUB_LOG_LEVEL=INFO
AI_HUB_MAX_RETRIES=3

Development

# Install dev dependencies
pip install -e ".[dev]"

# Run tests
pytest tests/

# Run linting
ruff check .
ruff format .

# Type checking
mypy src/ai_hub

# Pre-commit hooks
pre-commit install
pre-commit run --all-files

For Different Audiences

👔 Recruiters

This repository demonstrates:

• Production-grade Python development practices
• Deep understanding of LLM APIs and architectures
• System design for AI applications
• Clean, maintainable, well-documented code

💼 Clients

Examples show capability in:

• Building reliable AI-powered features
• Integrating multiple AI providers
• Implementing evaluation and quality control
• Production deployment patterns

👩‍💻 Developers

Learn from:

• Modern Python packaging (pyproject.toml, ruff)
• Provider abstraction patterns
• Error handling and retry strategies
• Testing LLM applications

License

MIT License - see LICENSE for details.

Contributing

Contributions welcome! Please read the contributing guidelines and submit a PR.

---

Built with modern Python practices and production-grade patterns.