At Fekra, I contributed to DIDASKIA, a platform for digitizing the internal processes of Institut de Formation Pédagogique de Lille. I developed features with Python and FastAPI, built web interfaces with Angular, and integrated REST APIs with PostgreSQL.
For DIDASKIA, I worked on RAG pipelines for document search, from ingestion and indexing to retrieval and contextualized answers. I also used Phoenix to monitor application quality and pipeline performance.
I helped industrialize the platform through GitLab CI/CD pipelines and containerized AI workloads deployed on Kubernetes in AWS. I also worked with Terraform, Helm, and HashiCorp Vault as part of the cloud and DevOps environment.
Earlier, at Fraikin Groupe, I contributed to COCKPIT, an internal application for tracking vehicle interventions and maintenance, using Java, Spring Boot, and Angular. I also built a personal multi-agent AI system with Microsoft AutoGen and Azure OpenAI for CSV analysis, PDF report generation, and Python script debugging.

