Jordan C.

Hubvisor is an AdTech platform enabling digital publishers to manage and optimize their programmatic monetization at scale.

The infrastructure is based on distributed systems processing massive real-time data streams (header bidding, log-level data, analytics), with high requirements in terms of performance, scalability, and reliability.

In this context, the main challenges involve designing high-performance data platforms, optimizing cloud costs, and implementing highly observable systems.

Responsibilities:

• Design and optimization of Kubernetes architectures on GKE

• Design and implementation of distributed data platforms (ClickHouse, Trino, Druid)

• Advanced tuning of clusters and data workloads

• Implementation of comprehensive observability (OpenTelemetry, metrics, debugging)

• Cloud cost optimization (spot instances, scheduling, storage)

• Deployment automation (Terraform, Helm)

• Participation in architecture and scaling decisions

Leadership:

• Structuring of Platform / DevOps practices

• Technical support and skill development for the team

• Contribution to strategic product and architecture decisions

Stack:

Kubernetes (GKE), Terraform, Helm, ClickHouse, Trino, Apache Druid, GCP, OpenTelemetry

Results:

• Reduction of cloud costs by 30-50% through resource and scheduling optimization

• Significant improvement in analytical query performance

• Stabilization of critical high-volume workloads

• Reduction of time-to-deploy through automation

• Implementation of observability enabling quick and efficient diagnostics

Armonik is an open-source HPC orchestrator designed to distribute and execute intensive computing workloads at scale on cloud and Kubernetes infrastructures.

The project aims to provide a scalable and multi-cloud platform for orchestrating parallel processing with strong performance, reliability, and elasticity constraints.

Responsibilities:

• Design and industrialization of Kubernetes deployments using Helm

• Development of multi-cloud Helm charts (AWS, hybrid environments)

• Packaging and integration for distribution via AWS Marketplace

• Implementation of DevOps best practices (CI/CD, standardization, automation)

• Contribution to the deployment architecture and platform scalability

• Collaboration with product and infrastructure teams

Stack:

Kubernetes, Helm, AWS, Docker, CI/CD

Results:

• Standardization and simplification of multi-cloud deployments

• Improved platform portability

• Accelerated deployment for clients

• Significant reduction in time-to-production

• Improved multi-cloud scalability and portability

As part of the Ségur du numérique en santé, participation in the DRIM-M project aimed at building a national secure network for sharing and consulting medical images (radiology and nuclear medicine).

The project is part of a critical environment with strong requirements in terms of security, scalability, and system reliability.

Responsibilities:

• Design and implementation of a Kubernetes platform for deploying and operating services

• Industrialization of deployments through CI/CD pipelines (AutoDevOps / NoOps)

• Complete automation of the delivery cycle (build, deployment, monitoring)

• Implementation of DevSecOps best practices

• Optimization of service performance and scalability

• Collaboration with product and technical teams

Stack:

Kubernetes, CI/CD, DevSecOps, Cloud, AutoDevOps, Docker

Results:

• Implementation of a robust Kubernetes platform for a national health system

• Reduction of time-to-deploy through complete automation (NoOps)

• Improved service resilience and availability

• Standardization of DevSecOps practices across the project