atef hafdhi

-Supported GenAI model training by evaluating and correcting AI responses in cybersecurity topics. -Designed prompt injection scenarios and assessed model performance using domain-specific criteria. -Applied expertise in penetration testing, OWASP Top Ten, and tools like Burp Suite and Nmap. -Contributed to enhancing AI understanding of real-world security threats and best practices.

-GitOps & Kubernetes: FluxCD management of 180+ namespaces across 2 production clusters (STG + PRD) for 20+ operator clients. Handled Istio service mesh, Kyverno policies, Vault + ExternalSecrets, cert-manager, ArgoCD, and full RBAC management. -K8s v3 - RKE2 Migration : Drove the architecture study and design for migrating from a legacy Kubernetes v3 monocluster to a multi-cluster RKE2 setup (5 clusters). Designed namespace strategy, ingress migration path (Istio → NGINX → Gateway API), FluxCD bootstrap process, Vault AppRole integration, and observability architecture. -CI/CD Platform : Built and maintained Piplib, a modular GitLab CI/CD shared library used across all Java, Node, and Helm projects in the org. Standardized delivery pipelines with quality gates, Docker builds, auto-deploy, and dynamic environment detection. -Helm Charts : Built versioned microservice Helm charts (meeseeks-charts) published to JFrog Artifactory, and maintained a reusable generic chart (common-helm-chart) powering all ITSF/RISF microservices. - AI Platform (KnowVA) : Led production-readiness assessment for a RAG/LLM system (Mistral-7B on H-100 GPUs, Qdrant vector DB, Celery, Django). Identified critical security/reliability gaps and delivered a full remediation and migration roadmap to Monaco Telecom infrastructure. -Mobile App Infrastructure : Managed CI/CD and Kubernetes infra for Galaxion, a React Native multi-brand mobile monorepo (Monaco + Digital). Fixed FluxCD deployment loops, API gateway routing failures, and pod-to-pod networking issues. -Network & Security : Calico GlobalNetworkPolicies, Istio AuthorizationPolicies, and Ansible automation for DNS, HAProxy, and VPN.

-Led an R&D project at Squad to automate penetration testing using AI. -Integrated Azure OpenAI services with GPT models for vulnerability scanning, web interaction automation, and report generation. -Optimized resource consumption and enhanced the functionality of large language models (LLMs). -Utilized tools such as Selenium WebDriver, OWASP ZAP, and the National Vulnerability Database (NVD) for comprehensive security testing. -Implemented prompt compression techniques using llmlingua to streamline AI interactions. -Developed a robust AI-driven pentesting framework, improving efficiency and accuracy in cybersecurity assessments.

Skills: Web Application Firewalls · Information Security · Artifactory · JFrog Software · Sonarqube · Jenkins · Gitlab · Dynatrace · Tableau · IT Audit · sast · SCA · Kubernetes

Participation in an audit mission within BNP Paribas, aiming to improve the security posture of the infrastructure and promote DevSecOps methodologies. The audit focused on the toolchain provided for users, with the objective of identifying security and regulatory compliance gaps, as well as the best practices to be applied.

Skills: Amazon Web Services (AWS) · Docker · OVH · PostgreSQL · Linux System Administration · System Administration · Ansible · Gitlab · Process Automation · Kubernetes · Jira · Maven · Oracle Database

-Create the company's DevSecOps CI/CD self-service for Hosting new projects (Gradle, Maven) and deploy them up to production. -Implement and automate Test Driven Security : analysis of tool licenses, dependencies(SCA),source code testing(SAST) , dynamic app testing(DAST), intrusion attacks, analysis of Docker images. -Handle Build and Deployment of artifacts into DEV, QA, and PROD hosted on-premises environment (Proxmox). -Create gitlab jobs for build,testing and deployments and creating Git repos for source code and setup branching & tagging strategy to maintain the source code in the GIT repository. - Manage containers using Docker ,automating deployment of images to self-hosted container registry.

-This project is created to get experience on Microservices With Netflix OSS. This is a simple project with simple business requirements. -Develop Web application for Travel agency using spring boot -Netflix Eureka is used for managing instances -Netflix Zuul to allow browser, mobile app, or other user interface to consume services from multiple hosts -Docker is used to make it easier to create, deploy, and run applications by using containers.

-Made use of maven for building,testing and deploying applications and written structured pom which can can be consumed by Jenkins. -Setup sonarqube projects and provide code analysis in different programming standards. -Deploy JAR applications to Nexus using Jenkins auto deployment. -Making use of wide range of plugins available to setup smooth developer friendly workflows. -Setup full CI/CD pipelines so that each commit a developer makes will go through standard process of software lifecycle and gets tested well enough before it can make it to the production. -Extensively worked on Jenkins to implement continuous integration (CI) and Continuous deployment (CD) processes -Worked on installation of Docker using Docker toolbox. Created custom Docker container images, tagging and pushing the images. -Deploying Docker images into AzureCloud Vms using Jenkins agents for each Stage (Staging, production)

-In the context of optimizing Unmanned Aerial vehicles (UAV) Base station, the goal is to deliver autonomous intelligent control algorithms to enhance deployment of drones and minimize energy consumption using artificial intelligence tools . -Study project environment, challenges and different approches/algorithms. -Submit a scientific sheet of conducted research . -Set up a tested and simulated algorithm for boosting energy efficiency(EE) .

-Study and understand different kpis and architecture of operator -Deploy ftp server for the operator to load logs -Develop a script with python able to scan alarms logs and provide comprehensible state of sites(eNodB) -Integrate the script in «Django» Framework and to be self-activating every 5 minutes -Develop a model with high accuracy to predict site’s state in the future