Kenyan AI Climate Forecast Platform

I want to turn the concept below into a production-ready platform that any Kenyan resident can open on a phone or browser and immediately see an accurate, easy-to-read forecast for rainfall, temperature, and wind speed. Behind the scenes the system should fuse ENSO indicators (especially the ONI index) with historical station data and real-time feeds, run them through Random Forest, Gradient Boosting, and Logistic Regression models, and serve county- and town-level predictions.

Scope of work
• Data pipeline – ingest, clean, and store global ENSO metrics, satellite observations, and local station readings in a way that is ready for modelling and always current.
• Model development – train, validate, and compare the three algorithms above; document feature importance and performance; expose the chosen ensemble through a REST/GraphQL API.
• Web + mobile front ends – one responsive website and companion mobile apps (Android & iOS) that call the API, display intuitive Grafana-style dashboards, and push real-time alerts.
• Risk analytics – highlight drought, flood, and other extreme-weather probabilities for all 47 counties.
• Subscription engine – let users pick counties or towns and receive SMS/e-mail/app notifications when thresholds are crossed.
• DevOps – containerised deployment (Docker/Kubernetes preferred), CI/CD setup, and monitoring so the service stays resilient during severe events.

Deliverables
1. Source-controlled code for data ingestion, modelling, and API.
2. Trained models with reproducibility documentation.
3. Responsive website and published Android/iOS apps.
4. Grafana dashboards wired to live data.
5. Alerting and subscription module.
6. Deployment scripts and runbook.

Acceptance criteria
– Forecasts for rainfall, temperature, and wind speed update at least hourly and achieve agreed-upon accuracy benchmarks for each county.
– Front ends load in under 3 s on 3G connections.
– Alerts arrive within 2 minutes of a trigger event.
– Full hand-over with documentation and a walk-through session.

If you have a proven record in climate data science, scalable back-end engineering, and cross-platform UI, I’m ready to review your approach and timeline.