Khaled ArsalaneDevOps & Platform Engineer

Now: DevOps & Platform Engineer at Devoteam

Khaled Arsalane

I design and run multi-cluster Kubernetes platforms and automate the infrastructure underneath them.

GitHubLinkedIn
See my journey
Portrait of Khaled Arsalane

About

I’m a DevOps & Platform Engineer at Devoteam. Most of my days are spent on Kubernetes: managing workloads across multiple clusters, automating how those clusters are provisioned and kept up to date, and turning that into a platform other teams can rely on. On the cloud side, I work on Scaleway and increasingly with Terraform.

Outside work, I run a 10-node k3s homelab. It hosts a few dozen self-hosted services and is my playground for certification prep and for trying new technologies and approaches, AIOps included. See how it’s built

Before that I did a PhD at IRISA (University of Rennes) on scalable data stream processing in heterogeneous environments. I built ScaleHub to provision geo-distributed Kubernetes testbeds, tuned Apache Flink for gains of up to 3.5×, and taught 168 hours of Cloud, Networks and Databases.

Earlier: a Kubernetes operator in Go for hardware-accelerated network functions at Orange, and flight software for a reusable-rocket demonstrator at CNES.

Stack

Cloud & orchestration
  • Kubernetes
  • Multi-cluster management
  • Kubernetes operators
  • Docker
  • Scaleway
  • k3s (10-node homelab)
Infrastructure & automation
  • Ansible
  • Infrastructure as Code
  • Terraform
  • CI/CD
  • Linux
  • Networking (SR-IOV, VNF)
  • AIOps
Languages
  • Go
  • Python
  • C
  • Bash
Research
  • Apache Flink
  • Stream processing
  • Performance analysis
  • Benchmarking
  • Scientific writing
Beyond the terminal
  • Photography
  • Self-hosting
  • Aviation & aerospace
  • Chess
  • Cycling

daily drivers

Journey

  1. Now

    1. Building a GitOps Cluster-as-a-Service platform: Kubernetes clusters declared in git and provisioned automatically across cloud providers.

      KubernetesMulti-clusterGitOpsCrossplaneTerraformScalewayPlatform Engineering

      Platform engineering on a GitOps Cluster-as-a-Service setup. Teams declare the clusters they need in git, and the platform provisions and manages them with Crossplane v2 and Terraform, with Scaleway among the target cloud providers. The work centres on multi-cluster management, automating the cluster lifecycle, and keeping day-2 operations reliable.

      Key Achievements

      • Build a GitOps Cluster-as-a-Service platform where Kubernetes clusters are declared in git and reconciled continuously
      • Orchestrate cluster provisioning with Crossplane v2 and Terraform across several cloud providers, Scaleway among them
      • Automate the cluster lifecycle across the fleet, from provisioning to day-2 operations

      Links

  2. 2025

    1. Developed ScaleHub, an infrastructure provisioning tool using IaC and Kubernetes for benchmarking data stream processing systems.

      ResearchDistributed SystemsDevOpsCloud ComputingInfrastructure as CodeApache FlinkKubernetesAnsiblePythonTeachingPublic Speaking

      See the project: ScaleHub

      Research on how stream processing engines such as Apache Flink scale when a cluster mixes different hardware and network conditions. Experiments ran on Grid5000 and combined HPC nodes with Raspberry Pi devices across several sites, all orchestrated by Kubernetes, alongside a teaching load in Cloud, Networks and Databases.

      Key Achievements

      • Built ScaleHub, an open-source tool that provisions geo-distributed Kubernetes testbeds and runs repeatable benchmarks on them
      • Tuned Apache Flink on heterogeneous clusters, for performance gains of up to 3.5×
      • Published at IC2E 2024 and SAC 2026, and taught 168 hours of Cloud, Networks and Databases to 200+ students

      Links

  3. 2022

    1. Final-year internship: implemented an on-demand L2/L3 traffic processing VNF with Kubernetes automation and hardware acceleration support.

      CloudInfrastructureNetworkVNFKubernetesOperator SDKGoHardware AccelerationFaucetSR-IOV

      The internship explored running network functions as Kubernetes workloads instead of dedicated appliances: a traffic processing function that can be deployed on demand, configured automatically, and still use the NIC’s hardware acceleration.

      Key Achievements

      • Developed a VNF that processes L2/L3 traffic on demand, with Faucet as the SDN controller
      • Wrote a Kubernetes operator in Go (Operator SDK) that deploys and configures the VNF automatically
      • Enabled hardware acceleration with SR-IOV on NVIDIA Mellanox NICs, giving VNF pods direct access to NIC virtual functions

      Links

  4. 2021

    1. Bachelor’s and Master’s internships on MINI APTERROS, a reusable-rocket demonstrator: avionics integration, flight software in C, and a ground-station HMI in Python.

      Software EngineeringEmbedded SystemsCPythonGNCPCBIMUBeagleBoneQt

      MINI APTERROS is a small reusable-rocket demonstrator built within CNES’s PERSEUS programme. Across two internships, worked on its avionics end to end: connecting the BeagleBone flight computer to its sensors and actuators, writing the software that flies it, and building the tools to follow and command it in real time.

      Key Achievements

      • Integrated the avionics: IMU, GPS and servo actuators wired to the BeagleBone flight computer through hand-soldered PCBs
      • Wrote the flight software in C, including the full guidance, navigation and control (GNC) logic and the PID loops driving the servos
      • Built a Qt/Python ground-station HMI for live telemetry and telecommand

      Links