GROPYUS berlin, germany
Senior Data Engineer aug 2022 — present
- Build Python and SQL ETL/ELT pipelines for production planning, element lifecycle tracking, manufacturing reporting, and operational monitoring.
- Orchestrate CDC, PostgreSQL, lakehouse, and graph database workflows with Dagster, Airflow, dbt, and DLT; release tested changes through Azure CI/CD.
- Implement anomaly detection for data pipelines and monitor them with Prometheus, Grafana, and operational alerts, supporting improvements to robot and manufacturing efficiency.
- Use Terraform, Docker, Kubernetes, and Unleash feature flags to support platform delivery and controlled releases.
- Redesigned Python graph ingestion using PostgreSQL metadata and checksums to skip unchanged subgraphs, improving ingestion speed by approximately 50%.
- Improved GraphDB ingestion, operations and monitoring, reducing cluster alerts from 300 to 5 per half-year.
- Resolved increased run volume and restrictive Linkerd timeouts to PostgreSQL, reducing Dagster pipeline failures from 10% to below 1%. Connected graph updates to automatic Dagster runs.
- Introduced dynamic buffers to select suitable inputs when data arrives late, keeping downstream outputs on schedule; added data lineage and custom alerts.
- Developed task-sequencing and movement-path algorithms for insulation-blowing and screw-fastening robots, designed to minimise travel and mission time.
- Prototyped natural-language querying and interpretation of building digital twins using LLMs. Use Claude Code and agentic coding workflows.
- Work with Helm and Azure DevOps alongside the existing platform delivery stack.