[ Case Study ]
FakeNews MLOps Pipeline
End-to-end misinformation detection pipeline with containerized serving for reproducible single-command deployment.
Built an end-to-end MLOps pipeline using TF-IDF + Logistic Regression trained on the ISOT dataset, served via Flask + Gunicorn, containerized with Docker for reproducible deployment.
Core Problem
ML models for fake news detection are often trained as one-off scripts without a reproducible, deployable pipeline, making them difficult to serve or iterate on.
Solution
Built an end-to-end MLOps pipeline using TF-IDF + Logistic Regression trained on the ISOT dataset, served via Flask + Gunicorn, containerized with Docker for reproducible deployment.
Outcome
Achieved approximately 98.5% test accuracy. Delivered a containerized, production-ready serving pipeline as a complete academic project.