Adnane Aabbar

Client: ASSURFIN - Automation of insurance claims processing. Development of a pipeline to automatically process and store emails and attachments in MongoDB. Use of AI models for OCR and object detection in images to generate relevant descriptions of attachments. Identification of reasons for rejection based on legal criteria via prompt engineering. Generation of rejection documents based on the reasons detected and associated case law. Deployment of the solution on Azure for continuous claims processing. Client: CNP Assurances - Automatic extraction of non-financial information from reports, articles, and external sources. Retrieval of available non-financial data (taxonomy, carbon emissions, etc.) from Euronext and Sustainability Disclosures, then SRNAV to retrieve the link to the ESG report (PDF). Exploration of the company's official website to locate the CSR report. Use of Azure AI Document Intelligence to accurately extract the necessary green taxonomy and carbon indicators.

- Realization of PoCs for key account clients. - Demos of the platform and vulgarisation of AI concepts during presentations with customer business profiles. - Development and optimization of an artificial intelligence platform: --> Deploy, manage and optimize models in production environments. --> Implement experiment tracking and monitoring of trained models using MLflow. --> Lead the redesign of the platform's NLP module (text classification, named entity recognition, etc.). --> Implement an OCR pipeline for PDF and image data using various techniques of extraction of relevant text information. --> Implementing a pipeline for data processing of audio data, destined for audio classification and other conversation AI use cases. --> Optimization of the data analysis and pre-processing module with new methods for cleaning and imputing missing data for various data types. --> Implement an advanced Retrieval Augmented Generation(RAG) pipeline for FAQ data and evaluating it using different metrics and methods. --> Development of APIs in Python (Flask) to perform data processing, training models and inference with trained models. --> Add new methods and models to the existing set of Machine Learning and Deep Learning models for all AI tasks. (Classification , Regression, Time series, Clustering, Anomaly detection, NLP...)

Teaching algorithmic and programming to students of all levels. Helping professionals make their career change by introducing them to computer science, progamming. Providing Data Science MSc students with guidance on their university assignments and projects. Main subjects: - Algorithmic - Programming (Python) - Databases (SQL) - Data Science

- Studies and developments on customer projects covering the entire life cycle of ERP, CRM, BPM, etc. - Contributions to the company's open-source software. Main technologies used: Axelor Open Platform, Java, PostgreSQL, Git, JPA/Hibernate, Google Guice.

Forecast of future imports on the moroccan territory : - Understanding of the business - Design of mathematical forecasting models for time series - Data analysis and preparation for the Machine Learning part (Pandas) - Development of models (Tensorflow, Keras, Sickit-learn) - Deployment of simulation models on a web interface (Django)

Building a brick in the decision-making system for PORTNET S.A which includes : - Functional study part(understanding of the business, and inventory of indicators, management rules, calculation methods). - Study of sources, development of SQL queries using Oracle 10g and SQL Developer. - Choice of the ETL tool and installation of the environment(Talend OS). - Development of ETL flows. - Modeling of the business layer. - Development of dashboards using Tableau.

Building an e-commerce platform for selling electronic goods via MERN Stack : - Analysis of the need and study of the existing, - Modeling and design of the database (MongoDB), - Construction of the backend (RESTful API) (Node.js), - Building a frontend (React.js)