Douraïd B.

Generative AI & Intelligent Agents

• Development of personal GPTs and APIs for intelligent agents (Agentic AI)

• Implementation of RAG pipelines (Azure Search, CosmosDB, OpenAI) to answer business queries from internal documents.

• Creation of Azure Function Skills to enrich cognitive search flows.

• Infrastructure as Code (Bicep), YAML CI/CD pipelines, Docker, continuous deployment on Azure.

Model Context Protocol (MCP) Server Development

• Design and implementation of MCP servers to expose business capabilities as tools consumable by LLM agents.

• Integration with Azure OpenAI and FastAPI to enable rich contextual interactions.

• Contribution to the agentic ecosystem: standardization of tool calls, session management, authentication.

Industrial Computer Vision

• Design of an anomaly and defect detection system based on PaDiM (patch distribution modeling) for quality inspection. • Experimentation on real data, optimization of accuracy and inference speed.

• Flame detection and localization using image processing (OpenCV, Python) for fire safety systems.

• AI Model Development: Designed and implemented machine learning architectures for license plate recognition, including supervised training and character assembly algorithms.

• Data Pipeline Management: Collected, cleaned, preprocessed, and augmented large datasets to ensure high-quality training inputs.

• User Interaction: Built and deployed a user-friendly interface to facilitate seamless interaction with the AI recognition system.

• Performance Optimization: Improved accuracy and robustness through iterative model refinement and data-driven experimentation

• API Development: Designed and implemented robust APIs to support system functionality and integration.

• Web Application Development: Built and maintained scalable web applications with a focus on performance and usability.

• Quality Assurance: Conducted unit and integration testing to ensure reliability, functionality, and code quality.