Abdelaziz Lemriss

Designed and developed Data Applications (Data Apps) focused on data entry, governance, and business user adoption. Built end-to-end solutions combining frontend interfaces (Streamlit / Dash / Gradio), business logic, and data pipelines within the Databricks ecosystem. Modeled and implemented robust, scalable, and governed data structures using Unity Catalog. Developed and orchestrated data ingestion, synchronization, and exposure pipelines across the Lakehouse architecture. Implemented advanced data security strategies, including Row-Level Security (RLS) with Lakebase (PostgreSQL). Delivered high-quality user-centric interfaces (UX/UI) to ensure strong adoption by business teams. Developed advanced Python components: data processing, internal APIs, and application logic. Industrialized data workflows with CI/CD, logging, monitoring, and observability frameworks. Produced detailed architecture documentation, data models, and engineering guidelines. Supported testing, deployment, and production release processes, ensuring reliability and scalability.

As part of an international program in the veterinary field, Royal Canin provides a digital platform used by veterinarians around the world during consultations. This platform allows veterinarians to enter patient information (cats and dogs), such as health status, weight, nutrition, dietary recommendations, and consultation follow-ups, with the aim of improving animal welfare and personalizing nutritional recommendations: Analysis and mapping of data sources and definition of tables required for dashboards. Design and development of ETL pipelines with Azure Data Factory, reducing the processing time for data from multiple sources. Optimization of performance on Azure Databricks, with significant improvements in the speed of computing jobs. Implementation of use cases as part of cost reduction for the client. Migration and adaptation of existing pipelines to Databricks (Unity Catalog). Design of analytical tables for Power BI consumption. PMO role in addition to the Data Engineer position. Close collaboration with business teams to define new tracking and functional improvement requirements for the platform and anticipate the data impacts of application developments. Active participation in project management using agile methodology, with 4-week sprints. Progress monitoring, stakeholder coordination, prioritization of data issues. Key role in communication between different teams, ensuring alignment between business needs, technical constraints, and project planning. Supervision of junior consultants.

Designed and developed an intelligent assistant based on RAG (Retrieval-Augmented Generation) to support internal auditors in accessing and analyzing large volumes of documentation. Built an end-to-end pipeline including document ingestion, chunking, embedding generation, and semantic retrieval. Implemented vector search and similarity matching to enable accurate and context-aware responses. Integrated LLMs to generate reliable, explainable answers grounded in internal knowledge bases (PDFs, reports, guidelines). Improved information retrieval efficiency, significantly reducing time spent searching across internal documents. Designed evaluation approaches to assess answer relevance, faithfulness, and context quality. Collaborated with business users to align the solution with real operational needs and ensure adoption. Contributed to defining best practices for Generative AI use cases in enterprise environments.

Project to migrate Chanel's data platform to Azure Fabrics. The mission consisted of proposing a suitable data architecture (bronze, silver, gold), developing the necessary ETL pipelines, and creating semantic models and Power BI dashboards: Scoping phase: designing the data architecture on Azure Fabrics (bronze, silver, gold). Developing transformation scripts and creating ETL pipelines in Azure Fabrics. Creating semantic models and Power BI dashboards. Collaborating with the data team to ensure the consistency and quality of data flows.

International project aimed at leveraging data from Danone purchasing platforms for different markets. The goal was to consolidate data from multiple sources, create a generalizable matching algorithm, and build transformation pipelines to feed business dashboards for marketing campaigns: Analysis and ingestion of data from different sources by market and rollout. Development of a matching algorithm to consolidate users between different sources. Design and implementation of ETL pipelines to transform data and create final tables. Application of machine learning models for use cases such as churn and customer segmentation.

Project at Webtec aimed at minimizing the gap between the expected delivery date and the actual delivery date of railway products. The goal was to accurately predict the delivery date for future orders in order to improve planning, reduce penalties for delays, and optimize revenue: Exploratory data analysis on SQL Server to identify patterns and logistical insights. Creation of a data transformation pipeline on SSIS to automate processing and prepare data for modeling. Development of machine learning models to predict exact delivery dates. Validation of models and evaluation of predictive performance. Design and implementation of Power BI dashboards to visualize KPIs and predictions for business teams.