Ayi NEDJIMI - Cybersecurity & AI Tools

100+ open-source cybersecurity, AI, and GPU computing tools

View on GitHub

HuggingFace Models and Datasets

Profile: huggingface.co/AYI-NEDJIMI

Back to Home


Overview

These models and datasets are designed for French-speaking cybersecurity professionals who need AI tools trained on domain-specific security content, compliance frameworks, and French regulatory requirements.


Fine-tuned Models

CyberSec-Assistant-3B

A 3 billion parameter model fine-tuned for cybersecurity tasks: threat analysis, incident response procedures, security policy writing, and vulnerability assessment. Trained on real-world incident reports, CVE descriptions, and MITRE ATT&CK content.

Use Cases: Automated threat report generation, SOC alert triage, security documentation drafting

View on HuggingFace


RGPD-Expert-1.5B

Specialized 1.5B parameter model for GDPR/RGPD compliance assessment, Data Protection Impact Assessment (DPIA) generation, and data protection advisory. Trained on official CNIL guidance, EU GDPR regulation text, and compliance templates.

Use Cases: DPIA drafting, data mapping questionnaires, DPO advisory support

View on HuggingFace


ISO27001-Expert-1.5B

Expert model for ISO/IEC 27001:2022 compliance auditing and information security management system (ISMS) implementation. Covers all 93 controls from Annex A with contextual implementation guidance.

Use Cases: ISMS gap analysis, control implementation guidance, audit checklist generation

View on HuggingFace


M365-Expert-v3

Microsoft 365 security configuration expert covering Microsoft Defender for Endpoint, Defender for Identity, Intune MDM/MAM, Azure AD Conditional Access, and Exchange Online Protection.

Use Cases: M365 security baseline assessment, Conditional Access policy design, Defender configuration review

View on HuggingFace


Datasets (50+)

French-language cybersecurity training datasets for fine-tuning LLMs on security topics:

Dataset Description Records
mitre-attack-fr MITRE ATT&CK techniques and tactics in French 1000+
ad-attacks-fr Active Directory attack techniques and detections 500+
forensics-windows-fr Windows digital forensics artifacts and analysis 800+
threat-hunting-soc-fr Threat hunting queries and SOC playbooks 600+
m365-security-fr Microsoft 365 security configurations and hardening 700+
cve-top100-fr Top 100 CVEs with contextual analysis in French 100+

And 44+ more specialized datasets covering: malware analysis, OSINT, cloud security, incident response playbooks, penetration testing methodologies, and French regulatory compliance.


Using the Datasets

from datasets import load_dataset

ds = load_dataset("AYI-NEDJIMI/mitre-attack-fr")
print(ds["train"][0])

from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("AYI-NEDJIMI/CyberSec-Assistant-3B")

Back to Home 2026 Ayi NEDJIMI