Deployable AI/ML Pipeline Development
Publicada el 2026-07-25
Descripción de la oferta
I need a production-ready machine-learning pipeline that can power three different initiatives at once—healthcare, fintech, and computer-vision analytics—without falling back into “research-only” mode. The datasets behind these projects are mixed: traditional MySQL tables, CSV exports, plus thousands of unstructured images and text snippets. Everything must flow through a single, well-orchestrated system. You will start by building robust data-ingestion layers for both structured and unstructured sources, then move on to model development in Python using TensorFlow, Keras, and Scikit-learn. On the vision side the end result should be an analytics-and-visualization module that can operate on live frames, while the tabular side focuses on supervised classification tasks for healthcare compliance and fintech risk scoring. The entire workflow has to be version-controlled, containerised, and exposed through either Flask or Django so I can drop it straight onto my existing servers. Deliverables • Reproducible codebase (Git) that includes data loaders, preprocessing, training, and inference scripts • Containerised API (Docker) with endpoints for predictions plus a lightweight web dashboard showing the computer-vision analytics and visualisations in real time • Orchestrated pipeline scripts for automated retraining and model tracking • Clear setup & deployment documentation, along with performance benchmarks on representative data The stack and integrations are negotiable as long as the solution stays portable, well-documented, and truly ready for production use.
Skills
- Machine Learning
- Django
- Docker
- MySQL
- Python
- Flask
- Keras
- PostgreSQL
- scikit-learn
- TensorFlow
- Compliance
- Computer Vision
- Git
Fuente original: freelancer