How I Built Yelena, My Custom RAG-Based AI Assistant

⚡ Yelena AI Assistant

Automated RAG-Based Knowledge Engine for MI Tech Arsenal

Yelena is a custom-built AI agent designed to act as a 24/7 technical assistant. Using Retrieval-Augmented Generation (RAG), she actively crawls over 90+ published blog posts, processes complex technical queries, and delivers precise, context-aware answers grounded entirely in published expertise.

---

🚀 Core Features

• Live Sitemap Sync: Automatically discovers and indexes new blog posts via sitemap.xml. No manual data entry required.
• Neural Search: Leverages Sentence Transformers to understand the meaning and intent behind questions, not just simple keywords.
• Floating Web Integration: Deployed as a persistent, mobile-responsive widget directly into the Blogger platform.
• Context-Aware Persona: Fully briefed on admin rules, hardware specs, and software workflows.

---

🛠️ Tech Stack & Hardware

Software Stack:

• Language: Python 3.10+
• LLM Engine: Google Gemini 2.5 Flash
• Vector Database: ChromaDB (In-Memory)
• Embeddings: all-MiniLM-L6-v2
• Frameworks: LangChain & Streamlit

Development & Testing Hardware:

• CPU: Intel Core i5-12400F
• GPU: AMD Radeon RX 6600 (8GB GDDR6)
• RAM: 16GB 6000MHz DDR5
• Mobile UI Testing: Moto G64 (Rooted, YAAP Custom ROM)

---

💻 Quick Install

Follow these steps to get Yelena up and running.

1. Clone the repository

[code lang="Bash"] git clone https://github.com/MI-Musanna/Yelena-AI.git cd Yelena-AI [/code]

2. Activate Virtual Environment

[code lang="Bash"] python -m venv ai_env [/code]

# For Windows:

[code lang="Bash"] ai_env\Scripts\activate [/code]

# For Mac/Linux:

[code lang="Bash"] source ai_env/bin/activate [/code]

3. Install Requirements

[code lang="Bash"] pip install -r requirements.txt [/code]

4. Launch Yelena

[code lang="Bash"] streamlit run app.py [/code]

---

👨‍💻 System Architect

Mahdi Islam
Lead RAG Architect & Founder of MI Tech Arsenal

My GitHub Portfolio