Skip to content
View michaelromagne's full-sized avatar

Block or report michaelromagne

Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
michaelromagne/README.md

👋

As a ML Engineer, I have contributed end-to-end to the productionization of multiple AI products at Ubisoft, GitGuardian, and Sanofi.

Sanofi (Dec 2024 - Present) | MLOps Engineer - GenAI & LLMOps

  • Development of an Unstructured Data Pipeline (OCR+VLM with Docling, AWS Textract, and Bedrock) deployed via Terraform module
  • Share LLMOps best practices for GenAI teams at Sanofi : Weave, LLM as Judge, GenAI Experiments, cost monitoring...
  • Stack: AWS Lambda, S3, ECR, Step Functions, Claude Sonnet, Amazon Nova Pro, Docling, HuggingFace, AWS Textract, PyMuPDF, Pinecone, W&B Weave

GitGuardian (Oct 2023 - Dec 2024) | Machine Learning Engineer

  • Built the company MLOps stack from scratch: GitLab CI, SkyPilot, DVC, Dagster, BentoML, Helm, ArgoCD
  • Fine-tuned and integrated NLP models (CodeBERTa) into the Secrets Detection Engine, reducing false positives by 5x
  • Stack: Transformers, PyTorch, FastAPI, ONNX Runtime, AWS EKS, Django, Celery, Kubernetes

Ubisoft (Feb 2021 - Oct 2023) | Machine Learning Engineer

  • End-to-end fraud detection project for e-commerce transactions (Ubisoft Connect and Steam)
  • Led research tasks (feature engineering, semi-supervised learning) and implemented MLOps best practices
  • Saved 5% of net sales (~4M€/year) compared to previous fraud detection product
  • Stack: XGBoost, DVC, ClearML, AWS Sagemaker, ECS, Kubernetes, Hadoop, Snowflake, Spark

The Scouting Arena

  • Web platform helping football fans discover and scout new players using advanced analytics
  • Features player statistics, visualizations, and scouting tools for a broad audience

Portfolio LinkedIn Malt

👨‍🔬 Skills

Programming: Python expert

Machine Learning: ML, NLP, GenAI, PyTorch, Transformers, Scikit-Learn, ONNX

Generative AI: OpenAI API, AWS Bedrock (Claude, Nova), HuggingFace, Langchain, Docling, W&B Weave

DevOps: AWS (Lambda, Step Functions, Batch, EKS, ECS, Sagemaker, S3, Bedrock), Kubernetes, Docker, GitLab CI, GitHub Actions, Helm, ArgoCD, Terraform

MLOps: W&B Weave, DVC, SkyPilot, BentoML, ClearML, Mlflow

Data Viz: Streamlit, Grafana, Tableau

Data Engineering: Temporal, Dagster, Airflow, Spark, Hadoop (HDFS, Hive), Snowflake

And Team Work, Being friendly with colleagues and Goal oriented 😄

Contact

Please contact me through Linkedin, Malt or email.

Pinned Loading

  1. treeverse/dvc treeverse/dvc Public

    🦉 Data Versioning and ML Experiments

    Python 15.2k 1.3k

  2. vibrantlabsai/ragas vibrantlabsai/ragas Public

    Supercharge Your LLM Application Evaluations 🚀

    Python 12k 1.2k

  3. wandb/weave wandb/weave Public

    Weave is a toolkit for developing AI-powered applications, built by Weights & Biases.

    Python 1k 140

  4. dataforgoodfr/batch11_e_cartomobile dataforgoodfr/batch11_e_cartomobile Public

    Encourager et planifier la mobilité électrique dans les territoires avec l’Open-Data

    Jupyter Notebook 6 4