Vinzenz Stampf

• • Prototyped (PoC) a GDPR-aligned RPO role–profile matching engine on Cloud Run with a lightweight Streamlit UI, using LLM-based semantic matching; authored the information security & data protection concept and aligned it with the CISO

• • Established engineering practices (coding standards, branching/PR reviews) and mentored teammates via pairing and design sessions within a 5–7 person team

retail)

• • Built a multi-tenant BigQuery DWH (Dagster ingestion + dbt transformations) across 4–7 tenants and ~5–10 providers; daily freshness, ~5–10 TB/day

• • Delivered a layered architecture (raw → marts), cust om data catalog, and unified cross-provider layer (ID mapping + fuzzy matching in Python)

• • Implemented and operated self-hosted Dagster on GKE (dagster-daemon + user-code deployments) using Helm and Workload Identity, supporting 100– 500 jobs/day and TB/day per tenant

• • Designed and operated hybrid connectivity: GCP HA VPN to AWS site-to-site with BGP routing and CIDR range planning/reservations, ensuring stable cross cloud network connectivity for platform workloads

• • Shipped a secure, versioned partner API using FastAPI on Cloud Run with OAuth authentication, enabling external integrations

• • Standardized infrastructure and delivery: Terraform for IaC, GitHub Actions for CI/CD, and automated rollouts across 2 environments aligned with Google Cloud Foundation Toolkit best practices

Large-scale analysis of non-relational data on distributed computing grid; Python/pandas ETL; scikit-learn; ML classifier; C++ contributions; Grafana dashboards for cluster monitoring