Ceyhun A.

I build data pipelines and analytics backends: Kafka and Flink streams, Airflow and Spark on Kubernetes, and ClickHouse with row-level security behind FastAPI and SSO. Since December 2024 I've done this remotely for Mediastream AG on ComplyGuard, a compliance analytics SaaS for regulated iGaming. That engagement started here on Upwork as a trial. What I built there includes: - Tenant isolation on ClickHouse at two layers (an API query guard and database row policies), covered by a 57-test suite - A Kafka Connect and PyFlink path that turned model scores into windowed KPIs in ClickHouse - Airflow on Kubernetes for two data platforms, including ML pipeline stages and model deployment automation - SSO with 2FA, role-based access and audit logging for an analytics app and Apache Superset Work I'm a good fit for: - Streaming into ClickHouse or PostgreSQL with Kafka, Kafka Connect or Flink - ClickHouse schema, query and multi-tenant security reviews - Moving Airflow or dbt pipelines onto Kubernetes, or cleaning them up - Embedding Superset dashboards behind SSO, with locked-down roles Public projects (in my portfolio below): a real-time review sentiment pipeline with Kafka, Spark, MongoDB and Elasticsearch, a multi-node ELK setup with index lifecycle tiers, and Airflow, Kafka and Cassandra pipelines, each with its code and a write-up. How I work: I'm in Istanbul (UTC+3), so I overlap with European working hours and US East Coast mornings. I write down decisions, send short async updates, and hand over a runbook with the code. Send me the problem and what "done" looks like, and I'll reply with how I'd approach it and a first estimate.