Machine Learning Engineer with 5 years of experience helping companies bring ML projects to production through robust MLOps practices.
ML, NLP, GenAI, PyTorch, AWS Bedrock, HuggingFace, ONNX
W&B Weave, Skypilot, DVC, BentoML, ClearML, Mlflow
Docker, Gitlab CI, Kubernetes, Helm, Argo CD, Terraform
Lambda, Step Functions, Batch, EKS, ECS, Sagemaker, S3, Bedrock
Python expert, Django, Celery, FastAPI, uv
Streamlit, Grafana, Kibana, Tableau
Temporal, Dagster, Airflow, Hadoop, Spark, Snowflake
I'm a Freelance Machine Learning Engineer with 5 years of experience helping companies bring ML projects to production through robust MLOps practices. I provide expert guidance in ML, NLP, and GenAI, crafting and deploying state-of-the-art models for complex challenges.
I worked at Ubisoft for three years deploying Machine Learning models to detect in-game fraudulent transactions. This experience taught me how to manage end-to-end Machine Learning projects, with in-depth exploration of Infrastructure, Data Engineering and MLOps matters. The project led to 5% of net sales savings, about 4 million euros per year.
Then, I joined GitGuardian as a Machine Learning Engineer, where I built the MLOps stack from scratch using GitLab CI, SkyPilot, DVC, Dagster, BentoML, Helm, and ArgoCD. I fine-tuned and integrated NLP models (CodeBERTa) into the Secrets Detection Engine, reducing false positives by 5x. I also developed automated secret leakage remediation using AST parsers and OpenAI API.
Following GitGuardian, I started working as a Freelance MLOps Engineer at Sanofi, focusing on GenAI and LLMOps. I developed an Unstructured Data Pipeline (OCR+VLM with Docling, AWS Textract, and Bedrock) deployed via Terraform, and I'm onboarding teams on W&B Weave for GenAI experiments and LLM monitoring. I also work on an AI Agents Gateway project for Sanofi.
In my free time, I'm passionate about Football and data analytics. I built The Scouting Arena, a platform helping football fans discover and scout new players using advanced analytics and visualizations.
If you want to have a talk, please contact me on Linkedin or at michael.romagne@gmail.com.
- Development of an Unstructured Data Pipeline (OCR+VLM with Docling, AWS Textract, and Bedrock, metadata generation, chunking, vectorization), deployed for Sanofi teams via a Terraform module.
- Share LLMOps best practices for GenAI teams at Sanofi : Weave, LLM as Judge, Experimentations...
- Work on defining an AI Agents Gateway for Sanofi.
- Built the MLOps stack from scratch: GitLab CI, SkyPilot, DVC pipelines, Dagster jobs, Streamlit, ONNX Runtime, BentoML, Helm, ArgoCD.
- Fine-tuned and integrated NLP models (CodeBERTa) into the Secrets Detection Engine, reducing false positives by 5x (Django, Celery, Kubernetes, AWS).
- Developed PoCs on automatic remediation for leaked secrets (OpenAI API, AST parsers and code formatting).
- End-to-end Fraud Detection project in e-commerce transactions (Ubi Connect and Steam).
Led Research tasks (Feature Engineering, Semi-supervised learning) and put in place MLOps best practices (DVC, remote jobs on K8s, ClearML, model inference on AWS).
The project led to 5% of net sales savings, about 4 millions euros per year, compared to the previous fraud detection product.
- Time Series forecasting on Acquisition, Retention, Monetization and Ubisoft servers vCPU usage. Trained and deployed Generalized Additive Models to improve forecasts.
Research on Digital Twins to optimize IoT Systems. Data Science, Simulation and Monitoring of IoT systems.
NLP on Orange mobile phone and internet boxes logs to predict churn and customer satisfaction.
