Ramzi Naifar

Project: Order Management System Context: System for managing airline orders within an e-commerce context. • Served as SRE for production and pre-production environments, managing production load and resolving L2/L3 incidents (e.g., OOM, blocked partitions) to ensure system stability. • Improved CI/CD reliability by 30% through root-cause investigation and targeted pipeline optimisations. • Architected pre-production and production migration strategies for MongoDB, Kafka and Event Mesh workloads as part of MultiPaaS migrations across 9 airline namespaces, including Air France, Finnair and Qatar Airways. • Developed andmaintained Kafka offset migration tooling adopted as the standard approach for subsequent airline migrations, enabling reliable cluster-to-cluster transitions with minimal downtime and 0 data loss. • Drove migration enablement through technical leadership, troubleshooting and coaching, enabling 4 engineers to independently execute migrations and supporting the transition toward a multi-PaaS architecture. • Established OpenShift governance: audit logs piped to Splunk with Teams alerts on manual changes across 30 namespaces spanning 4 clusters, including customer-facing environments. • Enabled single-PR multi-environment promotions with Teamsnotifications, reducing manual operations by 80%. • Implemented a platform-wide CI/CD metrics extraction and aggregation solution aligned with DORA metrics, with advanced Kibana dashboards covering the 5 pipelines’ reliability and performance. • Designed the CI/CD, delivery packaging and phased migration strategy for the Travel Capability initiative, enabling shared service adoption across multiple business units. • Ledaproofof concept to decompose monolithic ArgoCD deployments into independent applications, reducing deployment scope and improving delivery performance and scalability. • Supported 3 parallel initiatives (OrMS,Travel Capability and Universal Distribution), providing CI/CD enablement,namespace provisioning, platform onboarding and deployment standards across OpenShift environments. • Investigated and optimised ArgoCD deployment performance, reducing deployment time from 45 min to 12min through improved parallelisation and identifying further concurrency improvements. Technologies: Argo CD, Argo Workflow, Artifactory, Azure, Bitbucket, Elasticsearch, GitHub, Grafana, Helm, Jenkins, Kafka/Strimzi, Kibana, Kubernetes, Maven, MongoDB, OpenShift, Power Automate, Python, Quarkus, ServiceNow, Solace Event Mesh, Splunk

Project: Airline Ticketing System Context: Airline ticket management system migration to Azure. • Scripted updates for cloud log access across 50 Git repositories, increasing NRE success rate by 27%. • Created test scripts and investigated issues for cloud-migrated features. • Introduced environment-aware switching, enabling scripts to adapt to feature changes and improving issue detection automatically. • Supported 5 PROD Migrations to the cloud, covering 421 airlines by monitoring, analysing system behaviour, and preparing/executing automated tests. • Analysed and optimised SQL queries in production, achieving a 99% reduction in execution time and processed data. • Built a Splunk dashboard for system assessment and performance insights. • Rolled out a centralised Jenkins solution managing jobs across 50 repositories. Technologies: Azure, Bash, Bitbucket, Confluence, Control-M, DBPLUS, Flexify, Git, Grafana, Jenkins, Kubernetes, Jira, Linux, MinIO, Octane, OpenShift, Oracle, Podman, Python, Splunk, SQL

Project: Vodafone TV STB & VSB Software Context: European middleware project delivering next-generation Set-Top Box and Video Sound Box solutions, integrating multiple OTT services. • Ensured setup, system administration, and maintenance of the Jenkins infrastructure and VSB/STB test boxes. • Developed CI/CD pipeline using Jenkins and Gitlab CI, improving build and deployment efficiency. • Integrated and customised software packages using Git and Buildroot, and maintained the compilation environment. • Managed GitLab branching strategy and daily release management for a 12-developer team, supporting continuous delivery across Scrum sprints. • Developed Bash scripts to automate binary security signing and Nexus artifact upload, streamlining the final release step and reducing execution time by 92%. • Created and configured Docker images for integration use cases. • Deployed testing and monitoring tools, and custom scripts using Ansible. • Configured Grafana/Prometheus dashboards monitoring 16 CI pipelines and 5 servers, enabling continuous visibility into build health and infrastructure status. Technologies: Ansible, Apache HTTP Server, Bash, Buildroot, Debian, Docker, Jenkins, Jira, Git, GitLab, GitLab CI, Grafana, Keytab, Make, Nexus, Prometheus, Python, SVN

Project: Smart Meter System Context: Platform for managing smart electricity meters for a major Austrian energy supplier. • Automated the smart meter deployment process in the management system and conducted performance testing and analysis for one million meters. – Implemented and maintained Python and Bash scripts leveraging Selenium, Requests, and multithreading to optimise performance tests. – Automated the preprocessing and display of results from logs and database data using SQL, Python, and Linux commands to facilitate performance bottleneck analysis. • Maintained, enhanced, and debugged microservice features. – Improved data generation by introducing dynamic date handling. – Developed Bash script for multi-group query functionality. Technologies: ActiveMQ, Asyncio, Bash, CentOS, Confluence, multithreading, Pandas, Paramiko, PostgreSQL, Python, Selenium, SoapUI, SQL

Project: Autonomous Driving Context: Development and integration of new features to the "PiCar-V" robot so it can autonomously drive. The project was realised using two methods: Approach 1 – Image Processing: • Developed a lane detection and steering control method using image processing techniques. • Collected and preprocessed traffic sign and pedestrian datasets. • Trained and evaluated detection models. • Deployed AI code on Raspberry Pi and validated robot car control functionality. Approach 2 – Deep Reinforcement Learning (DRL) • Built a simulator using Unity for realistic training environments. • Collected data from both the simulator and the physical robot. • Tuned and trained an Unsupervised Image-to-Image Translation Network (UNIT) model for Raspberry Pi compatibility, enabling image reconstruction from simulation to reality and vice versa. • Integrated the UNIT model, trained and evaluated multiple DRL algorithms in simulation, and deployed the best-performing one on real hardware. Technologies: C#, Deep Reinforcement Learning, GAN, Intel NCS 2, Linux, OpenCV, Python, PyTorch, Raspberry Pi, Stable Baselines, TensorFlow, Unity, Variational Autoencoder.