<- all projects
Machine learning as a service
Proposition built on open-source tooling that delivers predictions from ML models as a service: client systems connected, predictions (such as churn and conversion) directly usable in business processes.
what I did
- Python apps for preprocessing, scoring and export, orchestrated with Airflow
- MLOps infrastructure with MLflow
- Deployment with Docker Compose on a Linux VM, managed through Portainer
- Grafana monitoring of application and model performance
stack
PythonAirflowMLflowDockerGrafana