Adam Proust

Establishment of a supplier repository offering scanner solutions for vehicles and/or AI solutions.

Quotation of solutions proposed by suppliers and establishment of quality criteria.

Technical study of a tailor-made solution for the end client including technical stack (minimum server configuration,

storage, AI model) as well as hardware (cameras, lighting, physical scanner arch structure).

Implementation of a POC with deep learning and computer vision models on open data images of damaged vehicles.

Full project monitoring and management.

Stack: Deep learning (CNN, segmentation, object detection), Python, Streamlit, Docker

Machine learning model for characterizing maintenance of military and civilian aircraft fleets using natural language processing (NLP) based on technician reports.

Statistical / probabilistic analysis to identify problematic equipment / aircraft.

Application work of generative AI models for automated generation of maintenance documentation for the RAFALE (France and Export, FR-EN).

Stack: Red Hat computing cloud, NLP, LLM, XML (S1000D format), Python

Implementation and deployment of a Python API on a Linux server used to query machine learning models developed by me and put into production.

Modeling and production deployment of a dynamic model to predict the charging time of electric vehicles per charging session and vehicle type.

Work on predictive maintenance of the electric charging station network. Containerization of frameworks and instances via Docker.

Stack: Automated Data/Modeling pipeline with Airflow, Docker (MLOPS), Python, FastAPI, Cassandra