✨ GenAI Hybrid App

🚀 A full-fledged Generative AI application that combines:

• 📝 Blog Generator (with customizable tones)
• 📄 AI Summarizer (with RAG & embeddings logging)
• 🤖 Q&A over documents (retrieval-augmented generation)
• 📜 History & Version Control (blogs, summaries, embeddings with timestamps)
• ⚙️ Custom Prompt Editor (prompt engineering playground)

Built with LangChain, OpenAI/Groq LLMs, and Streamlit to showcase advanced Generative AI Engineering skills.

"I built a Generative AI hybrid application that combines blog generation, document summarization, and Q&A over documents using a RAG pipeline. It has a Streamlit frontend where users can generate blogs with different tones, summarize uploaded files, and even ask questions about documents through retrieval-augmented generation. I also added history tracking, embeddings logging, and a custom prompt editor to showcase prompt engineering skills. The project demonstrates end-to-end GenAI engineering, including LLM orchestration with LangChain, embeddings with ChromaDB, and frontend integration."

---

🛠️ Tech Stack

Frontend

• Streamlit → Interactive UI with multiple sections, file uploads, and live prompt editing.

LLM & GenAI

• LangChain → Orchestration framework for LLM, embeddings, and RAG pipelines.
• LLM → Configurable (OpenAI GPT, Groq, or other providers).
• Prompt Engineering → Editable system prompts for blogs, summaries, and Q&A.

Embeddings & Vector Store

• Embeddings → Created for uploaded documents & stored for retrieval.
• ChromaDB → Local vector database for document Q&A.

File Handling & Persistence

• JSON storage for blogs & summaries history .
• Embeddings logged into summarizer_embeddings.txt (for transparency & debugging).

---

🚀 Features

1️⃣ Blog Generator

• Input a topic + choose tone (formal, casual, marketing, academic).
• Generates human-like blogs using LLM.
• Saves each blog in blogs.json with timestamp.

2️⃣ AI Summarizer

• Upload PDF or TXT files.
• Choose summary style: concise, detailed, or bullet_points.
• Automatically generates embeddings for file chunks.
• Stores results in summaries.json + embeddings in summarizer_embeddings.txt.

3️⃣ Q&A over Documents (RAG)

• Upload a document → stored in ChromaDB as embeddings.
• Ask questions about the document → Answer is generated using retrieval + LLM .
• Demonstrates Retrieval-Augmented Generation .

4️⃣ History & Version Control

• View all past blogs & summaries in a structured format.
• Each entry shows timestamp, tone/style, content, and embeddings reference .
• Highlights state management & persistence (important for production apps).

5️⃣ Custom Prompt Editor (Prompt Engineering)

• Built-in Prompt Editor UI in Streamlit.
• Edit system prompts for:
  • Blog Generator
  • Summarizer
  • Q&A
• Showcase prompt engineering mastery (a must-have GenAI skill).

---

📂 Project Structure

├── app.py                  # Main Streamlit app
├── modules/
│   ├── blog_generator.py   # Blog generation logic
│   ├── summarizer.py       # File summarization + embeddings
│   ├── rag_qa.py           # Document Q&A with ChromaDB
│   ├── history.py          # JSON-based history persistence
│   ├── prompts.py          # Default system prompts (editable)
│   └── utils.py            # Helper functions
├── blogs.json              # Stores blog history
├── summaries.json          # Stores summaries history
├── summarizer_embeddings.txt # Logs generated embeddings
├── requirements.txt
└── README.md

---

⚙️ Installation & Setup

1. Clone the repo

git clone https://github.com/your-username/genai-hybrid-app.git
cd genai-hybrid-app

2. Create and activate a virtual environment

python -m venv venv
source venv/bin/activate   # Mac/Linux
venv\Scripts\activate      # Windows

3. Install dependencies

pip install -r requirements.txt

4. Add your API keys in a .env file

GOOGLE_API_KEY=your_openai_key
# or Groq, Gemini, etc.

5. Run the app

streamlit run app.py

---

📸 Screenshots

• Blog Generator UI

  

  

• AI Summarizer with styles

  

  

• Q&A over documents (RAG)

  

• History Viewer

  

• Prompt Editor (Prompt Engineering Playground)

  

---

🎯 Why this project is Resume-Worthy?

✅ End-to-End GenAI System : Not just a toy project – covers text generation, summarization, RAG, embeddings, and history mgmt.

✅ Demonstrates RAG : Employers look for RAG pipeline understanding → you showcase it via Q&A + summarizer.

✅ Prompt Engineering : Custom prompt editor UI is a unique highlight.

✅ Persistence : JSON + embeddings logging → production-style state management.

✅ Full-Stack AI Engineering : Combines LLMs, vector DB, orchestration, and a clean frontend.

✅ Scalable : Can easily plug in any LLM provider (OpenAI, Groq, Gemini, Anthropic).

---

🔮 Future Enhancements

• Support for multi-file Q&A (knowledge base).
• Deploy on Vercel/Streamlit Cloud .
• Integrate Pinecone/Weaviate for scalable embeddings storage.
• Add user authentication & personalized history.
• Export blogs/summaries as PDF/Word files .

---

👨‍💻 Author

Ashutosh Raj Gupta – AI Engineer | MERN + GenAI Developer