elharith azzouzi idrissi

• Migration of an industrial multi-client predictive maintenance platform (airlines) from on-premise servers to GCP, covering a broad scope of fleets and aircraft systems. • Design and deployment of ML pipelines for anomaly detection and failure prediction by fleet and by system, in order to anticipate failures and reduce maintenance costs. • Implementation of an event-driven architecture on GCP: Cloud Run Services triggered by Pub/Sub, dynamic scheduling of Cloud Run Jobs, and real-time feeding of BigQuery. • Development of a monitoring and alerting system for Cloud Run Jobs executions, enabling real-time supervision and rapid anomaly detection. • Industrialization of infrastructure as code via Terraform: automated provisioning of Cloud Run Services, Cloud Run Jobs, and Pub/Sub flows. • Implementation of a CI/CD pipeline on GitHub Actions to ensure reliable deployments on a critical platform. • Development of a GenAI & OCR pipeline for document processing: page classification via LLMs (LangChain), text extraction through OCR, and key-field inference using dynamic prompts. • Deployment of this pipeline on GCP through a Cloud Run Services (scheduling) and Cloud Run Jobs (execution) architecture.

• Deployment and optimization of predictive models to enhance data & AI products within Carrefour. • Design and deployment of a RAG chatbot on GCP using Cloud Run: end-to-end ingestion pipeline (chunking, embedding, indexing), vector databases, semantic caching with Firestore, LangChain orchestration with Gemini, agents, and advanced prompting. • Implementation of OCR and NLP pipelines for automatic document extraction and structuring. • Use of LLMs for automatic document classification and key-field extraction. • Leadership of a couponing project featuring a personalized recommendation system. • Production-grade delivery of AI models (versioning, deployment, testing, monitoring), development of batch pipelines on GCP (Dataflow, Kafka, Dataform), orchestration with Vertex AI and Airflow, CI/CD, Docker, and application monitoring. • Setup of performance testing and continuous validation to ensure scalability, reliability, and quality of deployed models. • Integration of advanced business rules ensuring budget compliance and robustness of AI systems. • Close collaboration with business and infrastructure teams to guarantee reliable, cross-functional, and industrialized integration of AI solutions.

• Design and development of forecasting models • Ensuring the proper functioning and improvement of the various pipelines used daily • Improvement of the existing Marketing Mix Model (automatic data retrieval, model architecture change ...) • Long-term studies and carrying out analyses that allow the Marketing Department / General Management to steer the company's activity based on the results provided • Optimization of the media plan for each type of insurance • Improvement of existing web applications • Identification of data science opportunities through various studies conducted • Migration of scripts to the cloud service used (Azure, AWS)

• Development of web applications for machine learning and data visualization. • Development of tools for textual analysis. • Construction of Marketing Mix Models. • Intervention on several customer segmentation topics. • Participation in the deployment of web applications.