Fabrice Ghysdael

• Implementation of a craftsmanship culture (TDD, BDD, clean code) in the teams (Paris, Chennai, Lisbon).

• POC and training on AI-assisted development (GPTOSS, RooCode).

• POC and migration to CI for E2E test automation with Cypress, Playwright, Robot Framework.

• POC and implementation of mutation testing (PIT Test) and stored procedure tests with UTPLSQL.

• DevOps migration from an internal solution to Gitops with ArgoCD for Kubernetes and to Ansible / XLRelease for VMs.

• Unblocking of CI tests, blocked for 1 year.

• High-performance CI (Jenkins / GitlabCI / ArgoCD / Kubernetes) with optimized pipelines for each application.

• 30% reduction in production bugs.

• Teams trained in AI-assisted development (20+ developers).

• MDM: architecture, design, and development for incorporating evolutions, implementation of GraphQL and Kafka, deployment optimization (Helm and Kubernetes), observability setup (Prometheus, Grafana).

• SSO: based on Keycloak, implementation of two-factor authentication (2FA).

• Leadtime: complete redesign in 3 months (DDD model, Postgres BDD schemas, microservices architecture, Java/Spring boot, GitlabCI CI/CD) of the application, which had been dysfunctional for 3 years. BDD and API performance tests via Jmeter, integrated into GitlabCI.

• API Key Management: based on Kong. V2 architecture (DDD model, bounded contexts, UI, software and system architecture, user workflows). Deployment on Kubernetes via Helm and GitlabCI.

• Greenfield development of a solution for backing up the MDM MongoDB database to AWS S3, triggered via Kubernetes CronJobs.

• Implementation of craft and clean code best practices within the team.

• Development and integration into CI (GitlabCI) of automated E2E tests (Cypress) for MDM and API Key Management.

• 50% acceleration of production deployments with Kubernetes.

• Functional delivery date prediction APIs with performance of 34ms at 1500 calls/s.

• 30% reduction in MDM production bugs thanks to automated tests, observability, and MongoDB backups restorable in minutes for bug reproduction.

• Teams trained in craft and clean code (5 developers).

• Design and architecture of the solution in DDD: model, bounded contexts, ubiquitous language with domain experts and the 5 developers of the team, in sync with the architecture department.

• Development of the BFF and participation in the greenfield development in Java Spring Boot with a microservices architecture based on Kafka (outbox pattern with PostgreSQL then Kafka stream).

• Implementation of CI with GitlabCI and CD with Gitops based on ArgoCD / Helm / Kubernetes (Google Cloud GKE).

• Development of automated test solutions (Robot Framework) in CI and an API stub solution. Mutation testing (PIT Test) to validate unit test quality.

• Development of an on-demand deployment solution (BFF, microservices, stub server, test Kafka, Robot Framework test automation) on local Kubernetes (Docker Desktop) to allow developers to work autonomously.

• Training the team on TDD / BDD / Craft and clean code best practices through katas and workshops.

• No technical debt or defects 6 months after production.

• High-performance business application: beneficiary management, standard and instant bank transfers.

• High-performance CI based on Gitops, TTM halved compared to initial estimates.

• Teams trained in TDD / BDD / Craft (5 developers).