Looking for a job

ADAM DESORMIERE

@adamdesormiere

AI / DataOps Engineer at Orange Business building RAG applications and cross-cloud data pipelines for bpost.

At Orange Business, on the bpost account, I built an in-house GPT-OSS / NVIDIA-Nemotron RAG application for sensitive-data reasoning. It supported two major refactorings of legacy data marts.

I also led two AWS / Azure data migrations covering 8M customers and built a GDPR-compliant ETL pipeline for 10M legacy records. When data corruption affected 100k+ customers, I led the response and built a recovery pipeline across Dataverse, Azure, and Oracle.

At Intel, I built ETL pipelines for math-equation datasets and trained and fine-tuned machine learning models for equation clustering. My PhD research focused on deep learning for Intel electronic simulations to reduce time to market.

Experience

Work history, roles, and key accomplishments

Current

AI / DataOps - LLM Engineering, Data Engineering

Current

Orange Business

Apr 2025 - Present (1 year 7 months)

Built an in-house GPT-OSS / NVIDIA-Nemotron RAG application for sensitive-data reasoning and prototyped a machine learning customer deduplication algorithm on AWS. Led cross-cloud AWS / Azure data migrations and a recovery pipeline for data corruption affecting 100k+ customers.

LLM RAG GPT OSS Nemotron AWS Azure Data Factory Synapse Glue ETL GDPR Dataverse Oracle Python SQL

Data Research Engineer

Intel

Apr 2020 - Jun 2024 (4 years 2 months)

Built ETL pipelines for data cleaning, preparation and augmentation of math equations datasets. Trained and fine-tuned machine learning models for equation clustering and proposed a search engine for simulation models using business metadata.

ETL Data Cleaning Data Preparation Data Augmentation Machine Learning Embeddings Similarity Analysis Python SQL

Data Science Intern

Safran Electronics & Defense

Mar 2019 - Sep 2019 (6 months)

Benchmarked CNNs for real-time detection of military vehicles and other targets. Trained and fine-tuned YOLO / SSD neural networks for real-time object detection at long range.

CNN YOLO SSD Real time object detection Deep Learning Python

Data Science Intern

CNRS / LP2N

Mar 2018 - Jun 2018 (3 months)

Implemented OpenPose for real-time human pose estimation and markerless motion tracking. Calibrated a multi-camera setup for 3D reconstruction of human motion.

OpenPose Human Pose Estimation Motion Tracking 3D Reconstruction Computer Vision Python

Education

Degrees, certifications, and relevant coursework

INSA Lyon (dispLAB) & École Centrale Nantes (LS2N)

PhD, Machine Learning & Industrial Engineering

PhD in Machine Learning & Industrial Engineering. Research on deep learning for Intel electronic simulations to reduce time to market.

Paris-Saclay University | Institut d'Optique (IOGS)

MSc, Mathematics, Physics & Engineering

MSc in Mathematics, Physics & Engineering. Research on computer vision neural networks for real-time military object detection and human pose estimation.