Content
<div align="center">
# AI Engineering Hub
### Your comprehensive resource for learning and building with AI
[](https://github.com/damn8daniel/ai-engineering-hub)
[](https://opensource.org/licenses/MIT)
[](http://makeapullrequest.com)
<p align="center">
<img src="https://img.shields.io/badge/LLMs-blue?style=for-the-badge&logo=openai&logoColor=white" alt="LLMs"/>
<img src="https://img.shields.io/badge/RAG-green?style=for-the-badge&logo=elasticsearch&logoColor=white" alt="RAG"/>
<img src="https://img.shields.io/badge/AI_Agents-purple?style=for-the-badge&logo=robot-framework&logoColor=white" alt="AI Agents"/>
<img src="https://img.shields.io/badge/MCP-orange?style=for-the-badge&logo=protocol-buffers&logoColor=white" alt="MCP"/>
<img src="https://img.shields.io/badge/LLMOps-red?style=for-the-badge&logo=mlflow&logoColor=white" alt="LLMOps"/>
<img src="https://img.shields.io/badge/Evaluations-teal?style=for-the-badge&logo=pytest&logoColor=white" alt="Evaluations"/>
</p>
</div>
---
## Why This Repo?
AI Engineering is advancing rapidly, and staying at the forefront requires both deep understanding and hands-on experience. Here, you will find:
- **93+ Production-Ready Projects** across all skill levels
- **In-depth tutorials** on LLMs, RAG, Agents, and more
- **Real-world AI agent applications**
- **Examples to implement, adapt, and scale** in your projects
Whether you're a beginner, practitioner, or researcher, this repo provides resources for all skill levels to experiment and succeed in AI engineering.
---
## Table of Contents
- [Getting Started](#-getting-started)
- [Projects by Difficulty](#projects-by-difficulty)
- [Beginner Projects (22)](#-beginner-projects)
- [Intermediate Projects (48)](#-intermediate-projects)
- [Advanced Projects (23)](#-advanced-projects)
- [AI Engineering Roadmap](#-ai-engineering-roadmap)
- [Contributing](#contributing)
- [License](#license)
---
## Getting Started
New to AI Engineering? Start here:
1. **Complete Beginners:** Check out the [AI Engineering Roadmap](./ai-engineering-roadmap) for a comprehensive learning path
2. **Learn the Basics:** Start with [Beginner Projects](#-beginner-projects) like OCR apps and simple RAG implementations
3. **Build Your Skills:** Move to [Intermediate Projects](#-intermediate-projects) with agents and complex workflows
4. **Master Advanced Concepts:** Tackle [Advanced Projects](#-advanced-projects) including fine-tuning and production systems
---
## Projects by Difficulty
### Beginner Projects
> Simple, self-contained projects to get started with AI engineering.
#### OCR & Vision
| Project | Description | Tech Stack |
|---------|-------------|------------|
| [LaTeX OCR with Llama](./LaTeX-OCR-with-Llama) | Convert LaTeX equation images to code | Llama 3.2 Vision, Streamlit |
| [Llama OCR](./llama-ocr) | 100% local OCR application | Llama 3.2, Streamlit, Ollama |
| [Gemma-3 OCR](./gemma3-ocr) | Local OCR with structured text extraction | Gemma-3, Ollama |
| [Qwen 2.5 OCR](./qwen-2.5VL-ocr) | Text extraction from images | Qwen 2.5 VL |
#### Chat Interfaces & UI
| Project | Description | Tech Stack |
|---------|-------------|------------|
| [Local ChatGPT with DeepSeek](./local-chatgpt-deepseek) | Mini-ChatGPT with visible reasoning | DeepSeek-R1, Chainlit |
| [Local ChatGPT with Llama](./local-chatgpt-llama) | ChatGPT clone with vision | Llama 3.2, Streamlit |
| [Local ChatGPT with Gemma 3](./local-chatgpt-gemma3) | Local chat interface | Gemma 3, Ollama |
| [DeepSeek Thinking UI](./deepseek-thinking-ui) | ChatGPT with visible chain-of-thought | DeepSeek-R1, React |
| [Qwen3 Thinking UI](./qwen3-thinking-ui) | Thinking UI with streaming | Qwen3:4B, Streamlit |
| [GPT-OSS Thinking UI](./gpt-oss-thinking-ui) | Open-source GPT with reasoning viz | GPT-OSS, React |
| [Streaming AI Chatbot](./streaming-ai-chatbot) | Real-time AI streaming chatbot | Motia Framework |
#### Basic RAG
| Project | Description | Tech Stack |
|---------|-------------|------------|
| [Simple RAG Workflow](./simple-rag-workflow) | Basic RAG pipeline | LlamaIndex, Ollama |
| [Document Chat RAG](./document-chat-rag) | Chat with your documents | Llama 3.3, LangChain |
| [Fastest RAG Stack](./fastest-rag-stack) | Optimized RAG pipeline | SambaNova, LlamaIndex, Qdrant |
| [GitHub RAG](./github-rag) | Chat with GitHub repositories | LlamaIndex, Ollama |
| [ModernBERT RAG](./modernbert-rag) | RAG with modern embeddings | ModernBERT, FAISS |
| [Llama 4 RAG](./llama-4-rag) | RAG powered by Llama 4 | Llama 4, LangChain |
#### Multimodal & Media
| Project | Description | Tech Stack |
|---------|-------------|------------|
| [Image Generation with Janus-Pro](./imagegen-janus-pro) | Local image generation | DeepSeek Janus-Pro 7B |
| [Video RAG with Gemini](./video-rag-gemini) | Chat with videos | Gemini AI, LangChain |
#### Other Tools
| Project | Description | Tech Stack |
|---------|-------------|------------|
| [Website to API with FireCrawl](./website-to-api-firecrawl) | Convert websites to structured APIs | FireCrawl, FastAPI |
| [AI News Generator](./ai-news-generator) | Automated news generation | CrewAI, Cohere |
| [Siamese Network](./siamese-network) | Digit similarity detection | PyTorch, MNIST |
---
### Intermediate Projects
> Multi-component systems, agentic workflows, and advanced features for experienced practitioners.
#### AI Agents & Workflows
| Project | Description | Tech Stack |
|---------|-------------|------------|
| [YouTube Trend Analysis](./youtube-trend-analysis) | Analyze YouTube trends | CrewAI, BrightData |
| [AutoGen Stock Analyst](./autogen-stock-analyst) | Advanced stock analyst agent | Microsoft AutoGen |
| [Agentic RAG](./agentic-rag) | RAG with document search and web fallback | LlamaIndex, Tavily |
| [Agentic RAG with DeepSeek](./agentic-rag-deepseek) | Enterprise agentic RAG | GroundX, DeepSeek |
| [Book Writer Flow](./book-writer-flow) | Automated book writing | CrewAI Flow |
| [Content Planner Flow](./content-planner-flow) | Content workflow automation | CrewAI Flow |
| [Brand Monitoring](./brand-monitoring) | Automated brand monitoring system | CrewAI, LangChain |
| [Hotel Booking Crew](./hotel-booking-crew) | Multi-agent hotel booking | DeepSeek-R1, CrewAI |
| [Deploy Agentic RAG](./deploy-agentic-rag) | Private Agentic RAG API | LitServe, LlamaIndex |
| [Zep Memory Assistant](./zep-memory-assistant) | AI Agent with human-like memory | Zep, LangChain |
| [Agent with MCP Memory](./agent-with-mcp-memory) | Agent with persistent graph memory | Graphiti, Opik |
| [ACP Code](./acp-code) | Agent Communication Protocol demo | ACP SDK |
| [Motia Content Creation](./motia-content-creation) | Social media content automation | Motia Framework |
#### Voice & Audio
| Project | Description | Tech Stack |
|---------|-------------|------------|
| [Real-time Voice Bot](./real-time-voicebot) | Travel guide voice bot | AssemblyAI, ElevenLabs |
| [RAG Voice Agent](./rag-voice-agent) | Real-time RAG voice agent | Cartesia, LlamaIndex |
| [Chat with Audios](./chat-with-audios) | RAG over audio files | Whisper, LangChain |
| [Audio Analysis Toolkit](./audio-analysis-toolkit) | Comprehensive audio analysis | AssemblyAI |
| [Multilingual Meeting Notes](./multilingual-meeting-notes) | Auto meeting notes with language detection | Whisper, GPT-4 |
#### Advanced RAG
| Project | Description | Tech Stack |
|---------|-------------|------------|
| [RAG with Dockling](./rag-with-dockling) | RAG over Excel with IBM's Dockling | IBM Dockling, LangChain |
| [Trustworthy RAG](./trustworthy-rag) | RAG over complex docs with TLM | Cleanlab TLM |
| [Fastest RAG with Milvus and Groq](./fastest-rag-milvus-groq) | Sub-15ms retrieval latency | Milvus, Groq |
| [Chat with Code](./chat-with-code) | Chat with codebase | Qwen3-Coder |
| [RAG SQL Router](./rag-sql-router) | Agent with RAG and SQL routing | LangChain, PostgreSQL |
#### Multimodal
| Project | Description | Tech Stack |
|---------|-------------|------------|
| [DeepSeek Multimodal RAG](./deepseek-multimodal-rag) | MultiModal RAG | DeepSeek-Janus-Pro |
| [ColiVara Website RAG](./colivara-website-rag) | MultiModal RAG for websites | ColiVara, DeepSeek |
| [Multimodal RAG with AssemblyAI](./multimodal-rag-assemblyai) | Audio + vector database + CrewAI | AssemblyAI, ChromaDB |
#### MCP - Model Context Protocol
| Project | Description | Tech Stack |
|---------|-------------|------------|
| [Cursor Linkup MCP](./cursor-linkup-mcp) | Custom MCP with deep web search | Linkup, Cursor |
| [EyeLevel MCP RAG](./eyelevel-mcp-rag) | MCP for RAG over complex docs | EyeLevel, Claude |
| [LlamaIndex MCP](./llamaindex-mcp) | Local MCP client | LlamaIndex, Ollama |
| [MCP Agentic RAG](./mcp-agentic-rag) | Agentic RAG via MCP | Claude, MCP SDK |
| [MCP Agentic RAG Firecrawl](./mcp-agentic-rag-firecrawl) | Web-aware agentic RAG | FireCrawl, MCP |
| [MCP Video RAG](./mcp-video-rag) | Video understanding via MCP | Gemini, MCP SDK |
| [MCP Voice Agent](./mcp-voice-agent) | Voice-controlled MCP agent | AssemblyAI, MCP |
| [SDV MCP](./sdv-mcp) | Synthetic Data Vault via MCP | SDV, MCP SDK |
| [KitOps MCP](./kitops-mcp) | ML model management via MCP | KitOps |
| [Stagehand x MCP-Use](./stagehand-mcp-use) | Web automation with MCP | Stagehand, Playwright |
#### Model Comparison & Evaluation
| Project | Description | Tech Stack |
|---------|-------------|------------|
| [Evaluation and Observability](./eval-and-observability) | LLM evaluation pipeline | CometML Opik |
| [Llama 4 vs DeepSeek-R1](./llama4-vs-deepseek-r1) | Model benchmark comparison | Ollama, Python |
| [Qwen3 vs DeepSeek-R1](./qwen3-vs-deepseek-r1) | Reasoning model comparison | Ollama |
| [O3 vs Claude Code](./o3-vs-claude-code) | Coding model comparison | OpenAI, Anthropic |
| [Sonnet4 vs O4](./sonnet4-vs-o4) | Claude vs GPT comparison | Anthropic, OpenAI |
| [Sonnet4 vs Qwen3-Coder](./sonnet4-vs-qwen3-coder) | Coding benchmark | Anthropic, Qwen |
| [Code Model Comparison](./code-model-comparison) | Multi-model code benchmark | Various |
| [GPT-OSS vs Qwen3](./gpt-oss-vs-qwen3) | Open-source model battle | GPT-OSS, Qwen3 |
---
### Advanced Projects
> Production-grade systems, fine-tuning, and cutting-edge AI implementations.
#### Fine-tuning & Model Development
| Project | Description | Tech Stack |
|---------|-------------|------------|
| [DeepSeek Fine-tuning](./deepseek-finetuning) | Fine-tune DeepSeek models | Unsloth, Ollama |
| [Build Reasoning Model](./build-reasoning-model) | Build DeepSeek-R1-like reasoning | GRPO, Unsloth |
| [Attention Is All You Need](./attention-impl) | Transformer from scratch | PyTorch |
#### Advanced Agent Systems
| Project | Description | Tech Stack |
|---------|-------------|------------|
| [NVIDIA NIM Demo](./nvidia-nim-demo) | CrewAI Flows + NVIDIA NIM | CrewAI, NVIDIA NIM |
| [Documentation Writer Flow](./documentation-writer-flow) | Automated doc generation | CrewAI Flow |
| [Multi-Agent Deep Researcher](./multi-agent-deep-researcher) | Multi-agent research system | MCP, CrewAI |
| [Multiplatform Deep Researcher](./multiplatform-deep-researcher) | Cross-platform research | BrightData, LangChain |
| [Web Browsing Agent](./web-browsing-agent) | AI web browsing automation | CrewAI, Stagehand |
| [Paralegal Agent Crew](./paralegal-agent-crew) | Legal RAG paralegal agent | CrewAI, LlamaIndex |
| [FireCrawl Agent](./firecrawl-agent) | Corrective RAG agent | FireCrawl, LangGraph |
| [Context Engineering Workflow](./context-engineering-workflow) | Production context pipeline | TensorLake, Zep |
| [Parlant Conversational Agent](./parlant-conversational-agent) | Compliance-driven chatbot | Parlant Framework |
| [Stock Portfolio Analysis Agent](./stock-portfolio-agent) | Full-stack portfolio analyzer | React, CrewAI |
| [Guidelines vs Traditional Prompt](./guidelines-vs-traditional-prompt) | Prompt engineering comparison | Various LLMs |
#### Advanced MCP & Infrastructure
| Project | Description | Tech Stack |
|---------|-------------|------------|
| [MindsDB MCP](./mindsdb-mcp) | Unified MCP for all data sources | MindsDB |
| [Financial Analyst DeepSeek](./financial-analyst-deepseek) | Financial analysis agent | DeepSeek, MCP |
| [Graphiti MCP](./graphiti-mcp) | Persistent memory with knowledge graphs | Zep Graphiti |
| [Pixeltable MCP](./pixeltable-mcp) | Multimodal data orchestration | Pixeltable |
| [Ultimate AI Assistant](./ultimate-ai-assistant) | Multi-MCP server orchestration | Claude, Multiple MCPs |
#### Production Systems
| Project | Description | Tech Stack |
|---------|-------------|------------|
| [GroundX Document Pipeline](./groundx-doc-pipeline) | Production document processing | GroundX API |
| [NotebookLM Clone](./notebooklm-clone) | RAG with citations and podcasts | LangChain, TTS |
---
## AI Engineering Roadmap
New to AI Engineering? Follow the [complete roadmap](./ai-engineering-roadmap) with 10 stages:
| Stage | Topic | Resource |
|-------|-------|----------|
| 1 | Master Python | Harvard CS50p |
| 2 | AI with Python | DeepLearning.AI |
| 3 | Maths for ML | Khan Academy |
| 4 | Understanding LLMs | 3Blue1Brown |
| 5 | LLM Research | Andrej Karpathy |
| 6 | AI Agents | Anthropic Guide |
| 7 | Applied AI | CrewAI on Coursera |
| 8 | AI Protocols (MCP) | MCP Guidebook |
| 9 | Project-based Learning | This Repo |
| 10 | Books | AI Engineering - Chip Huyen |
```
Foundation ━━▶ Master Python ━━▶ AI with Python
│
Mathematics ━━▶ Maths for ML ◀━━━━━━━━━━┘
│
Understanding ━━▶ Understanding LLMs ━━▶ LLM Research
│
Application ━━▶ AI Agents ━━▶ Applied AI ━━▶ MCP
│
Mastery ━━▶ Project-based Learning ━━▶ Books
```
---
## Tech Stack Overview
| Category | Technologies |
|----------|-------------|
| **LLMs** | Llama 4, DeepSeek-R1, Gemma 3, Qwen 3, GPT-4, Claude |
| **RAG** | LlamaIndex, LangChain, Qdrant, Milvus, ChromaDB, FAISS |
| **Agents** | CrewAI, AutoGen, LangGraph, Swarm |
| **MCP** | MCP SDK (Python/TypeScript), Claude, Cursor |
| **Voice** | AssemblyAI, Cartesia, ElevenLabs, Whisper |
| **Fine-tuning** | Unsloth, PEFT, LoRA, QLoRA |
| **Frontend** | Streamlit, Chainlit, React, Gradio |
| **Infrastructure** | Ollama, LitServe, Docker, FastAPI |
---
## Contributing
Contributions are welcome! Here's how:
1. Fork this repository
2. Create a feature branch (`git checkout -b feature/amazing-project`)
3. Add your project in the appropriate difficulty folder
4. Include a README.md with setup instructions
5. Submit a Pull Request
### Project README Template
Each project should include:
- Description and use case
- Architecture diagram (if applicable)
- Prerequisites and setup instructions
- Step-by-step usage guide
- Example outputs/screenshots
---
## License
This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.
---
<div align="center">
**If you find this helpful, please give it a star!**
</div>
