FreeBSD ARM64 Industrial Edge AI Platform

An advanced E2E (End-to-End) platform built to demonstrate cutting-edge Data Science, AI / MLOps, and Embedded Security (RedTeam).

This project simulates a secure industrial edge device running on FreeBSD ARM64 that collects telemetry (vibration, temperature). It transports data across the network to a Containerized Rocky Linux Data Science Backend running an enterprise React dashboard and live Machine Learning pipelines.

Core Portfolio Capabilities Demonstrated

1. Embedded C & FreeBSD ARM64

• Writing bare-metal C applications (sensor_reader, telemetry_daemon) optimized for the ARM64 architecture.
• Automated orchestrations using QEMU User-Networking.

2. Data Science & MLOps

• Isolation Forest Anomaly Detection: Unsupervised machine learning models deployed locally to catch drifting metrics that human rules can't catch.
• Automated SQLite Persistence: Real-time ELT ingestion architecture managed entirely via Python FastAPI.
• Enterprise AI Dashboard: High-grade React (Vite) Glassmorphism dashboard leveraging Recharts for time-frequency data analysis.

3. RedTeam AI & Security Operations (SOC)

• Adversarial Machine Learning (White-Box Evasion): An AI attacks the AI. A fuzzer script (adversarial_evasion.py) loads the defensive Isolation Forest model and iterates mathematically to craft catastrophic physical bounds (e.g., 75°C temperature) that fool the classification boundary into predicting it as "Normal" telemetry.
• Protocol Fuzzing & Spoofing: Simulating physical attacker intervention rewriting telemetry payloads (sensor_spoof.py).
• Data Poisoning / Injection: Bypassing embedded defenses (telemetry_injection.py).
• Live Attack Feeds: A fully integrated visual feed triggering when ML rules categorize data anomalies.

Setup & E2E Testing

Running this project brings up the entire Docker pipeline and QEMU emulator out-of-the-box.

1. Ensure qemu-system-aarch64, qemu-efi-aarch64 and docker-compose are installed.

2. Download the FreeBSD QEMU disk image (not stored in git — 2.1 GB):

   cd qemu && ./download_image.sh

   The image (*.qcow2) is excluded from the repository due to its size. It is fetched directly from the official FreeBSD mirrors.

3. Run the master operator:

   chmod +x all_in_one.sh
   ./all_in_one.sh

4. Open the Premium SOC Dashboard at http://localhost:3000.

5. Simulate Data Science Anomalies and RedTeam AI evasion algorithms using the scripts in scripts/:

   chmod +x scripts/run_redteam.sh
   ./scripts/run_redteam.sh

   Or manually with the redteam/ python scripts.

Git LFS — Media Files

This repository uses Git LFS to store videos, audio, and music files (*.mp4, *.wav, etc.). Before cloning or pulling, make sure Git LFS is installed:

# Ubuntu/Debian
sudo apt-get install git-lfs
git lfs install

# macOS
brew install git-lfs
git lfs install

After a fresh clone, LFS objects are fetched automatically. If files appear as pointer text instead of binary content, run git lfs pull.

For exhaustive documentation, read docs/sdd/SDD.md and docs/arch/architecture.md.