AI repository intelligence

FR-AI / ANALYSIS

Why it is worth attention

Developed during an internship at DRDO, it offers a practical end-to-end ML pipeline with real-time predictions via live demo, comparing DNN and baseline models on standard cybersecurity datasets.

Who it is for

Cybersecurity researchers and analysts
Data scientists exploring ML for threat detection
Students learning applied ML in security contexts
Organizations seeking a demoable ML-based IDS

Use cases

Upload network traffic CSV files to detect malicious connections (DoS, Probe, R2L, U2R)
Classify executable file features as benign or malware
Compare Deep Neural Network vs Random Forest performance on the same data
Educational demonstration of ML pipeline from preprocessing to deployment

Strengths

Integrates two independent ML pipelines (network intrusion + malware) in a single app
Provides a live, accessible demo with both frontend and API endpoints
Includes baseline comparison (Random Forest) for performance benchmarking
Covers complete ML workflow: data preprocessing, model training, evaluation metrics, and deployment

Considerations

Currently relies on static datasets (NSL-KDD) rather than real-time packet capture
No authentication or user management; suited for demo/research, not production
Future improvements listed (e.g., explainable AI, multi-class malware) indicate current gaps

README quick start

🛡️ AI-Based Cyber Threat Detection Framework

AI-powered cybersecurity platform for Network Intrusion Detection and Malware Classification

Developed during my internship at DRDO

🌐 Live Demo: https://cyber-threat-detection-backend.vercel.app/

⚙️ Backend API: https://cyber-threat-detection-backend-205p.onrender.com/

📖 Overview

Cyber threats continue to evolve rapidly, making traditional signature-based security solutions insufficient for detecting modern attacks. This project presents an AI-Based Cyber Threat Detection Framework that leverages Machine Learning to automatically identify malicious network activity and classify potentially harmful files.

The framework combines two independent machine learning pipelines into a single web application:

🌐 Network Intrusion Detection System (NIDS)
🦠 Malware Detection System

Both models are deployed behind a FastAPI backend and accessed through an interactive React dashboard, enabling users to upload datasets and receive real-time predictions.

🚨 Problem Statement

Organizations generate enormous amounts of network traffic every day.

Among millions of legitimate connections, attackers attempt to perform activities such as:

Denial of Service (DoS)
Network Probing
Unauthorized Access
Remote-to-Local Attacks
User-to-Root Exploits
Malware Distribution

Traditional antivirus software and Intrusion Detection Systems primarily rely on signature-based detection, which struggles to identify:

Zero-day attacks
Unknown malware variants
Modified attack patterns
Previously unseen threats

Machine Learning provides a data-driven approach by learning patterns from historical attack data instead of depending solely on predefined rules.

🎯 Objectives

This project aims to:

Detect malicious network traffic using Machine Learning
Classify files as Benign or Malware
Compare Deep Neural Networks against Random Forest baselines
Provide real-time predictions through REST APIs
Deploy an end-to-end AI-powered cybersecurity solution

🏗️ System Architecture

                     +----------------------+
                     |    User Uploads CSV  |
                     +----------+-----------+
                                |
                                v
                    +-----------------------+

Related repositories

Similar projects matched by category, topics, and programming language.

MoonshotAI

Featured

Kimi-K3

Kimi K3 is an open-weight, 2.8T-parameter native multimodal agentic model with a 1M-token context window, designed for frontier coding, knowledge work, and reasoning tasks.

AI & Machine LearningAI Agents

3,348

lopopolo

Featured

harness-engineering

Harness Engineering is a methodology for improving coding agent outputs by carefully crafting the environment around them—providing curated context, tools, and executable constraints that encode an organization’s nonfunctional requirements and cumulative lessons.

AI & Machine LearningAI Agents

2,390

slvDev

Featured

esp32-ai

A 28.9 million parameter language model runs on an $8 ESP32-S3 microcontroller entirely on-device, generating simple stories at about 9.5 tokens per second.

AI & Machine LearningLarge Language Models

1,960