Abdallah Alaa Eddine Djamai

o Processing of various unstructured documents into structured data and integration into a conversational intelligent assistant agent for a law firm using OCR techniques, document parsing libraries, and on-premise language models on HuggingFace, deployed on AWS cloud, Docker containerization, and Kubernetes orchestration.

o Management of cloud infrastructure on Azure for an AI application at Caterpillar (Bergerat Monnoyeur IT team, under the direction of Nicolas HAMON, agile/SCRUM methodology), for intelligent document search with a Weaviate vector database and GenAI (Generative AI) RAG techniques with GPT-4o LLM models and Azure AI Foundry services. Infrastructure managed in Infrastructure as Code and automated deployments via CI/CD.

o Deployment and integration of an application for generating reports augmented by AI analyses, and automation of report sending to clients at Caterpillar.

o Implementation of intelligent KPI calculations at Point of Sale from raw data in SQL and Python, automation of the daily data processing pipeline (ETL/Spark) on a Linode machine and AWS cloud, with a Node.js server.

o Management of matrix and differential calculations for an installation wear forecasting application at SUEZ, with regression and ML models (Flask and FastAPI).

o Creation and deployment of an intelligent "Matching" application for matching housing candidates with available housing at RIVP using FastAPI and Flask, and deployment of the solution on-site and internal data platform.

o Creation, administration, and development of intelligent agents on Azure and Microsoft 365.

o Code audit and implementation of evolutions to various existing AI projects, documentation, code refactoring and optimization, AWS and Azure infra management (Terraform, CI/CD, MLOps Git).

End-of-studies internship at the secondary cycle laboratory, project "Real-Time On-Board Counting System in Crowded Scenes".

o Project management using agile development methods.

o Implementation of an AI-based crowd counting system on the NVIDIA Jetson nano platform with the YOLO model. Management of issues generated by the embedded system and the limitation of resources required for real-time processing and calculations.

o Implementation of unit tests to ensure the success of large-scale deployments.

o Deployment of Apache Kafka for transmitting results to a supervision server.

o Technical interventions with the internal systems team on the communication infrastructure and IT systems and software of affiliated companies.