Satellite Data AI Pilot Site (Bulgaria) – end-to-end workflow demo & validation

SOFIA SAT CLUB offers a pilot environment in Bulgaria to implement and validate AI-enabled Earth Observation workflows, from data ingestion to user-facing outputs. We can support the full chain: satellite data acquisition (primarily open data sources), preprocessing, model training/inference, validation, and publishing results as maps, simple dashboards, or reports for non-technical stakeholders (schools, municipalities, NGOs).
Our club runs hands-on “AI & data processing workshops using open-source tools” and includes satellite data analysis activities, which we can translate into a structured pilot task inside a project.

Scope / activities:

• Define 1–2 pilot use cases (e.g., vegetation status, land cover change, drought indicators, wildfire-related mapping, urban heat, flood susceptibility)

• Implement data pipeline (download/ingest → preprocessing → model → post-processing) using open tools

• Establish validation protocol (basic ground truth, cross-check with reference datasets, uncertainty reporting)

• Produce user-friendly outputs (QGIS layers, web map tiles, PDF briefing, simple dashboard)

• Run an end-user feedback loop (students/teachers/local stakeholders) and iterate

Expected inputs from consortium:

• Target use case + required outputs

• Preferred AI stack / baseline model (if any)

• Access to any project-specific datasets/tools

Deliverables we can own:

• Pilot workflow specification (inputs/outputs, architecture, reproducibility notes)

• Pilot dataset package (processed sample + metadata)

• Evaluation report (metrics + qualitative feedback)

• Demonstration assets (maps/dashboard + short explainer)

Typical effort / size: 1–3 person-months (can start minimal and scale)
Keywords: EO, AI pipeline, validation, reproducibility, open-source, pilot, mapping

We are happy to join first as a small task partner and scale our contribution once the consortium structure is confirmed.