Let's meet!

I'm Wiem Abbassi,
AI & Cybersecurity Engineer

Building intelligent systems that are secure, scalable, and human-centric

LLMs • RAG • Threat Intelligence • SOC Pipelines • ML • Python

My Works Download Resume
wiem@portfolio > wiem_abbassi.py
OPEN FOR ROLES
@dataclass
class WiemAbbassi:
role: str = "AI & Cybersecurity Engineer"
degree: str = "Master's in Cybersecurity @ ISI"
expertise: list = [
"LLM Defense-in-Depth & Gateway Architecture",
"Threat Intelligence & SOC Pipeline Automation",
"Multi-Agent RAG Systems & Machine Learning"
]
honors: str = "🏆 1st Place — CyberHorizon Hackathon"
status: str = "🟢 Available for AI & Cyber Opportunities"
def mission(self) -> str:
return "Building intelligent systems that are secure, scalable & human-centric."
Portfolio

Featured projects

A selection of projects spanning AI, cybersecurity, and full-stack development.

Enterprise LLM Security Gateway

Enterprise LLM Security Gateway (Defense-in-Depth)

Asynchronous multi-layered reverse proxy securing enterprise LLMs against prompt injection, jailbreaks, and PII leaks.

Skills:LLM SecurityFastAPIDeBERTaLlama GuardPresidioLoRA
GitHub ↗  ·  ▼ Click for details
CyberHorizon

CyberHorizon: SOC-Ready Threat Intel Pipeline

1st place Automated threat intel pipeline with human-in-the-loop adaptability.

Skills:PythonMISPThreat IntelAnomaly DetectionML
GitHub ↗  ·  ▼ Click for details
Threat Hunting Query Generator

Threat Hunting Query Generator via AI

Translate natural language into validated Splunk, KQL, and Elasticsearch queries using local LLMs.

Skills:LLMOllamaSplunkKQLStreamlit
GitHub ↗  ·  ▼ Click for details
Tunisian Darija AI Assistant

Tunisian Darija Agricultural AI Assistant

Voice-based Q&A in Tunisian Arabic for olive growers with disease classification and RAG safeguards.

Skills:RAGCNNVoice AILangChain
GitHub ↗  ·  ▼ Click for details
OWASP Top 10 LLM

OWASP Top 10 for LLM Applications

Hands-on red teaming of LLMs with Mistral 7B, prompt injection, and OWASP-guided mitigation strategies.

Skills:LLM SecurityOWASPPrompt InjectionMistral 7B
▼ Click for details
About Me

Turning complex problems into intelligent solutions

10+
Projects Done
4
Certifications
2+
Years Learning

Core Expertise

Machine LearningDeep LearningNLPComputer VisionLLMs & GenAIPrompt EngineeringRAGLangChainAI AgentsThreat IntelligenceAnomaly DetectionPythonMISPSplunk / KQLMITRE ATT&CKPower BIPandas / NumPyGit / GitHub

I'm Wiem Abbassi, a Computer Science graduate from Higher Institute of Computer Science with a deep passion for AI, cybersecurity, and building systems that make a real impact. My work spans LLM-powered security tools, machine learning, and full-stack development.

I specialize in building AI-powered security solutions from threat intel pipelines and LLM-based detection systems to RAG applications and multi-agent architectures. Every project is driven by a commitment to real-world impact and continuous learning.

Email
wiemabbassi14@gmail.com
Phone
+216 20 09 60 70
Location
Tunis, Tunisia
Resume

Education & experience

Computer Science graduate with hands-on experience in AI engineering, cybersecurity, and business intelligence.

My education

Sept 2025 — 2027

Master in Cybersecurity

currently at Higher Institute of Computer Science

Master's program in Cybersecurity at Higher Institute of Computer Science (ISI), focusing on advanced security concepts, threat analysis, and defensive strategies.
Sept 2022 — June 2025

Bachelor in Computer Science

Degree by Higher Institute of Computer Science

Bachelor's degree in Computer Science at Higher Institute of Computer Science (ISI), covering core computer science fundamentals.

My experience

Jul 2026 — Aug 2026

AI Security Engineer

in the RFC agency

Engineered an intelligent security gateway to protect enterprise LLMs against adversarial threats and enforce compliance in real time. Architected a multi-tier defense pipeline that mitigated prompt injection and jailbreak attacks pre-inference, slashing threat response latency by ~96%. Implemented comprehensive sensitive data protection and behavioral anomaly detection to ensure secure, privacy-first AI deployments.
Feb 2025 — May 2025

Artificial Intelligence Engineer

in the RFC agency

Designed and deployed a RAG-based solution with Microsoft Sentinel to automate incident investigation. Evaluated multiple LLMs for threat analysis and developed automated SOAR playbooks for incident response workflows.
Aug 2024

Business Intelligence Analyst

in the MAVISION agency

Cleaned and pre-processed raw datasets to ensure data quality. Designed interactive Power BI dashboards to communicate findings and implemented metrics to support business insights.
Since 2023

Media Member

at Securinets ISI

Active member in the media department. Designed visual edits, Photoshop compositions, and video editing for events, workshops, and social media promotions.
Since 2022

HR Department Lead

at Freeways ISI

Active member since 2022. Took responsibility of the HR department in 2023, managing recruitment processes, onboarding, and team coordination across the club.
Achievements

Competitions & awards

Competitive programming achievements and academic distinctions.

2026 CyberHorizon
Won 1st place — pre-event project
Designed and developped a SOC-ready pipeline that automates intelligent threat intel enrichment, scores data, detects poisoning, pushes to MISP, and implements human-in-the-loop for better adaptability pre-event.
2023 TCPC
Ranked 39th out of 100 teams
Participated in the Tunisian Collegiate Programming Contest and solved algorithmic challenges in data structures, graph theory and dynamic programming.
2023 ISI Contest
hosted by Freeways ISI
Took part in a programming contest tackling algorithmic and problem-solving challenges organized by Freeways ISI.
2023 Code Her Ways
hosted by IEEE x Freeways
Participated in a coding competition focused on creative problem-solving and collaborative innovation.
CyberHorizon CyberHorizon TCPC TCPC ISI Contest Code Her Ways CyberHorizon CyberHorizon TCPC TCPC ISI Contest Code Her Ways
Certifications

Licenses & certifications

Professional certifications from industry leaders in AI, cloud, and data science.

Anthropic
March 2026
Claude Code in Action
Anthropic
View credential
 NVIDIA
April 2026
Building LLM Applications With Prompt Engineering
NVIDIA
View credential
 Google
February 2026
Google Cybersecurity Professional
Google
View credential
 Microsoft
November 2024
Azure AI Fundamentals (AI-900)
Microsoft
View credential
Blog

Thoughts & writings

Deep dives into AI security, LLMs, and cybersecurity — published on Medium.

OWASP Top 10 for LLMs

OWASP Top 10 for LLMs Explained: 10 Risks That Can Break Your AI App

A hands-on breakdown of every OWASP risk for LLM applications — from prompt injection and training data poisoning to insecure output handling — with real-world attack scenarios and mitigation strategies.

LLM SecurityOWASPRed Teaming
📖 Medium  ·  Read article ↗  ·  10 min read
Enterprise LLM Security Gateway Architecture

Beyond System Prompts: How I Built an Enterprise Multi-Layer LLM Security Gateway

A deep technical walkthrough of building a 15-subsystem defense-in-depth security gateway — covering DeBERTa-v3 prompt injection detection, PII masking with Presidio, Llama Guard evaluation, and LoRA retraining workflows.

AI EngineeringDefense-in-DepthFastAPILoRA
📖 Medium  ·  Read article ↗  ·  14 min read
Coming Soon Article in progress — publishing soon on Medium

AI SecOps & Observability: Tracing, Anomaly Detection, and Autonomous Retraining for LLMs

Exploring production-grade LLM observability — from OpenTelemetry distributed tracing and behavioral anomaly detection to autonomous retraining pipelines that keep models secure and reliable at scale.

AI SecOpsObservabilityMLOps
Coming Soon
View all articles on Medium ↗
Contact

Let's make something awesome together!

Have a project, question, or just want to say hi? Send me a message!

wiemabbassi14@gmail.com

+216 20 09 60 70

in GH @