AI Engineer & Arabic NLP Researcher
ENSIA - École Nationale Supérieure d'Intelligence artificielle Algeria

Building production-grade NLP systems for 400M+ Arabic speakers. Focused on RAG architectures, competitive programming, and ML security. Currently researching Arabic morphological tokenization and training Algeria's national programming olympiad team.

class SamirGuenchi:
    def __init__(self):
        self.role = "AI Engineer & Arabic NLP Researcher"
        self.institution = "ENSIA"
        self.location = "Algeria"
        self.focus = [
            "Arabic NLP",
            "RAG Systems", 
            "Competitive Programming",
            "ML Security"
        ]
        
    def current_work(self):
        return {
            "research": "Arabic RAG architectures & morphological tokenization",
            "coaching": "National Programming Olympiad - Algeria",
            "building": "Production-grade NLP systems for RTL languages",
            "learning": "Advanced AI/ML at ENSIA",
            "security": "ML-powered threat detection systems"
        }
    
    def philosophy(self):
        return "Code that ships > Code that sits in notebooks"

• Arabic Natural Language Processing
• Retrieval-Augmented Generation (RAG) Systems
• Competitive Programming & Algorithm Design
• Machine Learning Security
• Medical AI & Healthcare Applications

Repository: Ministry-Regulation

Production NLP system for Arabic government documents. Implements semantic search over 500+ page PDFs using RAG architecture.

Technical Implementation:

• Python + LangChain framework
• Custom Arabic embeddings and tokenization
• Vector databases for semantic retrieval
• RTL text processing pipeline

Problem Solved: Transformed information retrieval from manual PDF searching to natural language queries in Arabic.

Key Innovation: First-class Arabic language support with proper morphological handling, not post-hoc English translation.

Repository: Qr_Analyzer

Mobile security application with ML-based phishing detection. Analyzes QR codes before execution to prevent attacks.

Technical Implementation:

• Flutter + Dart cross-platform development
• ML classification models for threat detection
• Real-time pattern recognition
• Security-first architecture

Problem Solved: Proactive threat detection before user compromise, not reactive damage control.

Key Innovation: Pre-scan analysis using machine learning to identify malicious patterns.

Repository: BUPA-Liver-Disorder-Analysis

End-to-end machine learning pipeline for medical diagnosis. Research-grade methodology with clinical application focus.

Technical Implementation:

• scikit-learn + pandas for data processing
• Statistical validation protocols
• Cross-validation and feature engineering
• Reproducible research methodology

Problem Solved: Healthcare AI requires reproducibility and rigorous validation, not just high accuracy scores.

Key Innovation: Production-ready medical ML with proper statistical rigor and clinical-grade evaluation.

Repository: Search_Algo

Interactive educational tool for algorithm visualization. Built for competitive programming students.

Technical Implementation:

• Python with visualization libraries
• Interactive UI/UX design
• Pathfinding algorithms (A*, Dijkstra, BFS, DFS)
• Real-time step-by-step execution

Problem Solved: Abstract algorithmic concepts become concrete through visual demonstration.

Key Innovation: Teaching-optimized design based on how competitive programmers actually learn.

View all projects: github.com/Samir-Guenchi

Context: 400 million Arabic speakers online. Most NLP infrastructure built for English, then poorly adapted.

Current Work:

• Proper tokenization for agglutinative morphology
• Embeddings that understand RTL context
• RAG systems with native diacritic handling
• Models trained on Arabic, not translated English

Goal: Infrastructure that treats Arabic as a first-class citizen in NLP systems.

Approach: Constraint-based thinking from competitive programming applies directly to production ML systems.

Applications:

• Transformer inference optimization
• Algorithm design for scale
• System building under real-world limitations
• Performance-critical code development

Philosophy: Code that ships beats code that sits in notebooks.

Institution: ENSIA (École Nationale Supérieure d'Informatique et d'Analyse des Systèmes)
Program: Computer Science & AI Engineering
Location: Algeria
Focus Areas: Arabic NLP, Machine Learning, Algorithm Design

Additional Role: National Programming Olympiad Coach

Research:

• Arabic RAG architectures
• Morphological tokenization systems
• Multilingual NLP infrastructure

Development:

• Production ML systems deployment
• Open-source Arabic language tools
• Security-focused ML applications

Education:

• Training national olympiad candidates
• Algorithm optimization techniques
• Competitive programming methodology

Seeking:

• Research collaborations in Arabic NLP
• ML/NLP engineering internships
• Open-source contribution opportunities
• Hackathons and ML competitions

Professional:

• LinkedIn: linkedin.com/in/guenchi-samir
• Email: samir.guenchi@ensia.edu.dz
• GitHub: github.com/Samir-Guenchi

Competitive Programming:

• Kaggle: kaggle.com/guenchisamir
• Codeforces: codeforces.com/profile/Guenchi_Samir_ia

Last Updated: January 2025