Skip to content
Yassir Halaoui
Lire en français

Competence

Cloud, Kubernetes and the pipeline that ships it

I treat delivery as part of the design. On my current mission, the pipeline that takes a commit to Azure Kubernetes also builds the image, scans it, checks quality and provisions the Kafka topics.

In short

Yassir Halaoui works across Azure (AKS, Blob Storage, Key Vault, Container Registry), AWS (EC2, S3, Lambda, RDS) and Google Cloud, with Kubernetes, Helm, Docker and Terraform. He automates delivery with Azure DevOps, GitLab CI, GitHub Actions and Jenkins, including SonarQube quality gates, Trivy scans with SBOMs and Datadog APM. He is an AWS Certified Cloud Practitioner.

Where I've used it

  1. Pluxee · 2025 to now

    Automated delivery from Azure DevOps to AKS: multi-stage Docker builds, Helm, Trivy scanning with an SBOM, a SonarQube quality gate, Terraform for Kafka topics and roles, and secrets in Azure Key Vault.

  2. Société Générale · 2022 to 2024

    Set up the Docker and Kubernetes development infrastructure, wrote the Kubernetes manifests and built the Jenkins CI/CD pipelines.

  3. Docaposte · 2024 to 2025

    Shipped a containerised Spring Boot and React platform through GitLab, with documents stored in Amazon S3.

  4. Consulting · 2022

    Ran Spring Batch workloads on Google Cloud Platform.

  5. Kempt, RevealTogether, TheQuizMaster

    Run on Vercel and Railway with GitHub Actions for CI and scheduled jobs, including the daily rebuild that publishes this site's queued articles.

Everything in this area

Azure
  • AKS
  • Blob Storage
  • Key Vault
  • Container Registry
  • Azure DevOps
  • API Management
AWS and Google Cloud
  • EC2
  • S3
  • Lambda
  • RDS
  • Google Cloud Platform
Containers and IaC
  • Kubernetes
  • Helm
  • Docker
  • Terraform
  • Nginx
CI/CD
  • Azure DevOps
  • GitLab CI
  • GitHub Actions
  • Jenkins
  • Husky
Quality and security
  • SonarQube
  • Trivy
  • SBOM
Observability
  • Datadog APM
Platforms
  • Vercel
  • Railway
  • Serverless

In depth

The pipeline is part of the product

A service that can't be deployed safely on a Tuesday afternoon isn't finished. Every service I ship has a pipeline that builds, tests, scans and deploys it the same way every time, and that pipeline lives in the repository next to the code it ships.

Supply-chain security that stays switched on

A Trivy scan and an SBOM on every image, a SonarQube gate focused on new code, and secrets that never touch the repository. Security that blocks every merge gets turned off within a month. Security aimed at new risk stays on.

Infrastructure as code, including the boring parts

Terraform for the Kafka topics and their access roles, not just the clusters. A topic that exists because someone clicked it into being ends up configured differently in every environment.

Questions

Can you set up CI/CD from Azure DevOps to AKS?

Yes, it's what I run on my current mission: multi-stage Docker builds, Helm charts, Trivy and SonarQube in the pipeline, Terraform for supporting resources and secrets from Key Vault.

Do you write Terraform?

Yes, including the resources people tend to leave out, such as Kafka topics and their access roles.

Are you AWS certified?

Yes, AWS Certified Cloud Practitioner (CLF-C02) since 2023. My deepest production experience is on Azure, with AWS and Google Cloud alongside.

Field notes

Shipped work

Related competences