🛡️ Athena AI

Enterprise-grade cybersecurity AI assistant powered by Qwen3.6-27B, Cyber LoRA, Retrieval-Augmented Generation (RAG), and a large-scale cybersecurity knowledge base.

---

📖 Overview

Athena AI is an advanced cybersecurity-focused Large Language Model (LLM) platform designed to assist security researchers, penetration testers, SOC analysts, students, and cybersecurity professionals.

The project combines:

• Qwen3.6-27B
• Cybersecurity-specific LoRA Fine-Tuning
• Qdrant Vector Database
• Retrieval-Augmented Generation (RAG)
• Large-scale Cybersecurity Knowledge Base

Athena AI is designed to provide intelligent cybersecurity assistance through contextual reasoning, semantic search, threat intelligence retrieval, vulnerability analysis, and security research support.

---

✨ Features

AI Capabilities

• Intelligent Cybersecurity Assistant
• Context-Aware Conversations
• Long-Term Memory System
• Threat Intelligence Analysis
• Vulnerability Research Assistance
• Security Report Generation
• Log Analysis
• Malware Research Support
• Incident Response Guidance

Retrieval Capabilities

• Semantic Search
• Vector Similarity Search
• Context Retrieval
• Large-Scale Knowledge Base Querying
• Real-Time Information Injection

---

🏗 System Architecture

                        ┌─────────────────┐
                        │      User       │
                        └────────┬────────┘
                                 │
                                 ▼
                   ┌────────────────────────┐
                   │   Athena AI Interface │
                   └────────────┬───────────┘
                                │
                                ▼
                   ┌────────────────────────┐
                   │ Memory & Context Layer │
                   └────────────┬───────────┘
                                │
          ┌─────────────────────┼─────────────────────┐
          │                     │                     │
          ▼                     ▼                     ▼

   ┌─────────────┐      ┌─────────────┐      ┌─────────────┐
   │   Qdrant    │      │ Cyber LoRA  │      │ Tool Layer  │
   │ Vector DB   │      │             │      │             │
   └──────┬──────┘      └──────┬──────┘      └──────┬──────┘
          │                    │                    │
          ▼                    ▼                    ▼

   ┌────────────────────────────────────────────────────┐
   │               Qwen3.6-27B Base Model              │
   └────────────────────────────────────────────────────┘
                                │
                                ▼
                        Final Response

---

🛠 Technology Stack

Artificial Intelligence

Component	Technology
Base Model	Qwen3.6-27B
Fine-Tuning	QLoRA
Framework	Unsloth
Quantization	4-Bit
Precision	BF16
Embeddings	BGE-Large

Backend

Component	Technology
Language	Python
API	Flask
Database	SQLite
Vector Database	Qdrant

---

🚀 Training Infrastructure

Training is performed using cloud GPU infrastructure.

Component	Specification
Provider	RunPod
GPU	NVIDIA H100 SXM
VRAM	80GB
CUDA	12.x
OS	Ubuntu 22.04
Framework	PyTorch
Fine-Tuning Library	Unsloth
Optimization	Flash Attention

---

🧠 Fine-Tuning Pipeline

Cybersecurity Dataset
          │
          ▼
Data Cleaning
          │
          ▼
Instruction Formatting
          │
          ▼
Tokenization
          │
          ▼
Qwen3.6-27B Base
          │
          ▼
QLoRA Fine-Tuning
          │
          ▼
Cybersecurity LoRA
          │
          ▼
Evaluation
          │
          ▼
Deployment

Fine-Tuning Configuration

Model: Qwen3.6-27B
Method: QLoRA
Quantization: 4-Bit
Precision: BF16
Optimizer: AdamW
Framework: Unsloth

---

🔍 Retrieval-Augmented Generation (RAG)

Athena AI uses Retrieval-Augmented Generation (RAG) to provide factual and up-to-date cybersecurity knowledge.

RAG Flow

User Query
     │
     ▼
Query Embedding
     │
     ▼
Qdrant Search
     │
     ▼
Top Relevant Chunks
     │
     ▼
Context Assembly
     │
     ▼
Qwen3.6-27B + Cyber LoRA
     │
     ▼
Generated Response

---

📚 Knowledge Base

Current target knowledge base:

95+ GB Cybersecurity Data

Sources:

• CVE Databases
• MITRE ATT&CK
• Security Advisories
• Threat Intelligence Reports
• Malware Analysis Reports
• Security Blogs
• Security Documentation
• Research Papers
• Incident Response Playbooks

---

🔗 Embedding Configuration

Embedding Model:

BAAI/bge-large-en-v1.5

Alternative:

nomic-embed-text

Configuration:

Chunk Size: 1000
Chunk Overlap: 150
Top K: 10
Distance Metric: Cosine

---

📂 Project Structure

Athena-AI/
│
├── app.py
├── requirements.txt
├── README.md
│
├── ai/
│   ├── llm.py
│   ├── memory.py
│   └── prompts/
│
├── rag/
│   ├── embeddings.py
│   ├── qdrant_client.py
│   └── retrieval.py
│
├── datasets/
│
├── tools/
│   ├── report_generator.py
│   ├── log_analyzer.py
│   └── file_manager.py
│
├── templates/
│
└── workspace/

---

⚙ Installation

git clone https://github.com/vrunalp199/Athena-AI

cd Athena-AI

python -m venv venv

Activate environment:

Linux:

source venv/bin/activate

Windows:

venv\Scripts\activate

Install dependencies:

pip install -r requirements.txt

---

🗄 Qdrant Setup

Using Docker:

docker run -p 6333:6333 \
-v $(pwd)/qdrant_storage:/qdrant/storage \
qdrant/qdrant

Verify:

http://localhost:6333/dashboard

---

📈 Embedding Generation

Install dependencies:

pip install sentence-transformers
pip install qdrant-client

Example:

from sentence_transformers import SentenceTransformer

model = SentenceTransformer(
    "BAAI/bge-large-en-v1.5"
)

embedding = model.encode(
    "Explain SQL Injection"
)

---

🌐 Deployment Architecture

Internet
    │
    ▼
Flask API
    │
    ▼
Athena AI Core
    │
    ├── Memory System
    │
    ├── Qdrant Vector DB
    │
    └── Qwen3.6-27B
            │
            ▼
      Cyber LoRA
            │
            ▼
        Response

---

📊 Performance Targets

Metric	Target
Model	Qwen3.6-27B
Dataset	95+ GB
Embeddings	GE-Large
Vector DB	Qdrant
Retrieval Latency	< 2 sec
Deployment	Docker
GPU	H100/A100

---

🛣 Roadmap

Phase 1

• Core AI Assistant
• Memory System
• Prompt Engineering

Phase 2

• Cybersecurity LoRA
• Advanced RAG
• Vector Search Optimization

Phase 3

• Multi-Agent Architecture
• Threat Intelligence Agent
• Malware Analysis Agent

Phase 4

• SIEM Integration
• Security Automation
• Enterprise Deployment

---

⚠ Disclaimer

Athena AI is intended for educational purposes, cybersecurity research, defensive security operations, and authorized security testing only.

Users are responsible for complying with all applicable laws and regulations.

---

📜 License

MIT License

---

👨‍💻 Author

Vrunal Patil
Computer Science Student
Cybersecurity Researcher, - AI Developer
⭐ If you find this project useful, consider starring the repository