All Projects

OncoClear

A production-ready MLOps pipeline for accurate breast cancer classification using machine learning.
Project
OncoClear
project id
oncoclear
Use this id to create a new project in ZenML
GITHUB REPOSITORY
https://github.com/zenml-io/zenml-projects/tree/main/oncoclear
Pipelines

Feature Engineering Pipeline

Pipeline that loads the Wisconsin Breast Cancer diagnostic dataset, performs preprocessing, and splits data into training and testing sets.

Training Pipeline

Pipeline that trains classification models (SGD and Random Forest) and evaluates them on test data, promoting the best performer to production.

Inference Pipeline

Pipeline that uses the production model to generate predictions on new data, leveraging the same preprocessing as during training.

Deployment Pipeline

Pipeline that deploys the production model as a FastAPI service, making it accessible via REST API with interactive Swagger documentation.

Recommended Stack

Stack Components

  • Orchestrator: default
  • Artifact Store: default
  • Step Operator: default
Details

OncoClear is an end-to-end MLOps solution that transforms raw diagnostic measurements into reliable cancer classification predictions. Built with ZenML's robust framework, it delivers enterprise-grade machine learning pipelines that can be deployed in both development and production environments.

What It Does

OncoClear delivers production-ready breast cancer classification through comprehensive MLOps pipelines. It processes medical diagnostic data with automatic model versioning, comparison, and promotion workflows, ensuring only the highest-performing models reach production.

How It Works

  • Loads and processes the Wisconsin Breast Cancer diagnostic dataset through automated cleaning
  • Engineers features optimized for medical diagnostic classification tasks
  • Trains multiple classification models (SGD and Random Forest) with comparative evaluation
  • Evaluates models using accuracy, precision, recall, and F1 score metrics
  • Promotes best-performing models to production automatically
  • Deploys models as containerized FastAPI services for real-time prediction
  • Provides clean API documentation for easy integration into healthcare systems
Gallery