Data Engineer

Own and harden the data-delivery and serving backbone of SurfaceScout, LiveEO's solution for monitoring changes around critical assets. You will make the path from detection to delivered customer insight reliable and repeatable, covering delivery-management and insight-generation services, post-processing steps, and the AWS and Prefect platform that runs them end to end.

What you'll do

• Co-own and extend the delivery-management service (SDMS) and the insight-generation and post-processing services, the path that turns model detections into delivered, customer-ready insights.
• Harden these into well-tested, documented services so they are no longer a single-person dependency.
• Own and maintain the AWS and Prefect delivery platform, meaning the Lambda-based glue, the orchestration flows and the cross-repo release, and keep it deterministic and repeatable.
• Manage the PostgreSQL and PostGIS data model behind delivery, including the schema and the migrations via Alembic.
• Be a reliable second owner for where the pipeline breaks. You unblock stuck orchestration runs, keep delivery SLAs met, and cut out manual steps.
• Automate QA across the serving path, covering schema and geometry integrity and coverage gaps, and maintain structured metadata and STAC entries so every delivery is traceable.

What you'll need

• Strong Python in production, not scripting-only.
• Building and running backend services with FastAPI (or similar), SQLAlchemy, Alembic and Pydantic.
• PostgreSQL, and spatial data (PostGIS, GeoAlchemy2 or equivalent) is a strong plus.
• AWS in production, mainly S3, Lambda, EC2 and IAM basics.
• Orchestration with Prefect, Airflow, Dagster or equivalent.
• Comfort handling geospatial raster and vector data (rasterio, geopandas, shapely, STAC, COG) at a level where you can move it reliably.
• Pragmatic delivery and a reliability and ops mindset, so you ship robust, well-tested services and keep them running.
• Ownership and verification mindset, so you validate outputs against reality and catch results that look right but are not, using judgment and not just code.

Nice to have

• Distributed compute with Ray or Anyscale.
• MLflow, or experiment and dataset versioning.
• Remote-sensing or geospatial foundations, or SAR exposure (Capella, Sentinel-1).
• Observability with structured logging, OpenTelemetry and alerting.
• Satellite provider APIs such as Capella, UP42, Planet and ICEYE.

What they offer

• Flexible working hours and hybrid work model.
• A collaborative and learning environment with frequent internal workshops, knowledge sharing sessions, journal clubs and hackathons.
• Office in Berlin Kreuzberg with free fruit, nuts and drinks.
• Urban Sports membership and BVG subsidy, corporate pension program.
• Potential to participate in the employee stock option program.
• A diverse and vibrant international environment of 30+ different nationalities.