Contact person: Real-Time AI Threat Detection

Phone:Show

Email:Show

Location: Riyadh, Saudi Arabia

Budget: Recommended by industry experts

Time to start: As soon as possible

Project description:
"I’m building an AI-driven security engine that runs natively on Windows machines and reacts the moment trouble appears. The system must:

• Monitor system activity continuously, recognise suspicious behaviour in real time and decide when the risk is high enough to act.
• Quarantine the affected host, process or file instantly to stop spread.
• Send a concise email alert (JSON or human-readable—your call) the second an event is confirmed so administrators can jump in.
• Block the malicious activity or connection path automatically, then keep watching to confirm the block held.

Model choice, training data and architecture are up to you as long as the end product meets these goals, installs cleanly on Windows 10/11 and Windows Server, and produces near-zero false positives after tuning.

Deliverables
1. Source code (Python, C#, Go or another mainstream language) with clear build instructions.
2. Trained model or ruleset plus a plan for continuous learning.
3. Installer or script that deploys the agent/service on Windows.
4. Admin guide covering configuration, email alert format and response workflow.
5. Demo session showing the system detect, isolate and alert on at least two live threat scenarios.

If you’ve previously worked with frameworks such as TensorFlow, PyTorch, scikit-learn, Microsoft [login to view URL] or similar for cybersecurity, let me know. Proven experience integrating low-latency detection with Windows internals (ETW, Sysmon, Powershell logging, WinAPI hooks, etc.) will stand out.

The sooner we can get a stable prototype running, the better—feel free to outline milestones for data gathering, model training and on-box testing when you respond." (client-provided description)