Contact person: Hybrid ML Model Development

Phone:Show

Email:Show

Location: Baramati, India

Budget: Recommended by industry experts

Time to start: As soon as possible

Project description:
"Our data lives in two worlds—a legacy on-prem database and several Google Cloud Storage buckets—and the goal is to turn those mixed sources into a reliable, production-ready machine-learning model. I need the full workflow: ingest the hybrid data, engineer features, experiment with appropriate algorithms, and then package, validate, and deploy the model so it can be called from downstream applications.

Key points
• Hybrid data sources: on-prem databases + Google Cloud Storage
• Primary objective: end-to-end machine-learning model development, not just exploration
• Robust, repeatable pipelines expected for extraction, transformation, training, and inference
• Clear documentation so my internal team can maintain and extend the solution

Deliverables
1. Codebase (Python preferred) that connects to both data environments, performs all preprocessing, and trains the model reproducibly
2. Automated pipeline scripts or workflow definitions (e.g., Airflow, Cloud Composer, or similar)
3. Containerised deployment artefacts and inference endpoint setup
4. README or run-book covering configuration, retraining steps, and dependency management

Acceptance criteria
• Pipeline runs successfully against a sample and full data set on both environments
• Model performance meets agreed-upon metrics defined during kickoff
• All components version-controlled and executable via a single command or CI/CD trigger

If you have solid experience juggling on-prem data with GCS and delivering production ML, I’m ready to dive in." (client-provided description)