datadrift

Testing out NannyML

Production-ready NLP classifier: fine-tuned DistilBERT across 19 e-commerce categories with FastAPI serving, real-time drift detection via Evidently AI, and a React analytics dashboard. Fully containerized with Docker + CI/CD.

Minimal MLOps regression skeleton (California Housing) with training pipeline, Evidently drift/performance report, FastAPI prediction service, Dockerized training/serving environments, ready for CI/CD extension

ML Project END to END

This repository hosts a machine learning tool for breast cancer classification, emphasizing model interpretability. It is deployed on Streamlit Cloud, with a PostgreSQL database for tracking results and data drift, and includes an automated retraining workflow.

DriftSiren is a production-grade platform for real-time data drift and quality monitoring. Built with Next.js, FastAPI, and Docker, it tracks feature drift, provides live alerts, and visualizes metrics on a sleek dashboard. Includes agent, APIs, Celery workers, and Kubernetes-ready setup.

Exploring what a real-time distributed model serving and data drift monitoring solution could look like within MLOps

Développer un modèle de prédiction pour scorer la solvabilité d'un client et le déployer pour une mise en production

Practical MLOps implementation showcasing MLflow for experiment tracking and model management, and Evidently for monitoring data drift and model health in production.

Fast, declarative data quality framework (Rust). Schema validation, drift detection, anomaly detection. Pandas/Polars/DuckDB support.