Louis Paret

At Mindee, we are building an infrastucture to return the text from documents. We are developing and hosting API and AI models (CPU / GPU). - Migration of the on-premise infrastructure to Kubernetes to improve the scalability, stability and the flexibility. We used EKS, Helm and Terraform. - Creation of an Internal Development Platform for ~10 developers based on Kubernetes, Helm and ArgoCD. - Automation of PostgresQL deployments on Kubernetes to use it in production. - Cloud Cost Optimization : 50% reduced (Saving plans, Spot instances, Resources usage monitoring, Autoscaling, Design optimization, Opencost...) -Deployments of many internal tools and middlewares using Helm (Harbor, Thanos, Redis, RabbitMQ, Sentry, Prometheus/Grafana, Tailscale)

I have left France to discover the world during 16 months with my wife. I made 16 countries in 4 continents, I learned to speak spanish and I perfected my english. The travel developed my sense of the organization, anticipation, my curiosity and my openness. I also worked for some compagnies as a DevOps Cloud Engineer Freelance during the travel.

We're managing an huge infrastructure with more than 6 billions of requests per day with a high SLA. I realised a lot of tasks/projects: - Creation of an HA infrastructure for Prometheus with Docker/Ansible/Jenkins. - Automation of the deployment of an API on more than 25 endpoints on different Cloud Provider (AWS/GCP/Azure/Vultr/Scaleway/OVH) with Jenkins, Terraform, Packer, Ansible and Docker. - Load balancing of all the API traffic from 1 DNS to the most relevant endpoint (Geolocation, availability) - Redesign of the integration and continuous deployment pipeline of an application based on PHP/Angular with Jenkins, Docker, AWS, Kubernetes (GKE) - Management of an Elasticsearch cluster with more than 200 nodes with ILM, different layers and cost optimization. - Redesign of the provisioning/deployment scripts to simplify the addition of Kafka and Flink scripts - On-call to keep a maximal availability. - Optimization of the scaling of the API to handle huge DDOS attacks (AWS Autoscaling, Packer)