At Inspira grupa, I introduced Apache Flink as a production stream-processing platform, replacing a legacy real-time processing system. I also led the greenfield development of an internal candidate-evaluation and inverse recommendation system, designing its Neo4j graph model and candidate-matching relationships.
I built the system’s production platform with FastAPI services and processing flows integrated with Kafka, Flink, Airflow, and Spark. I developed workflows for historical data migrations and daily interaction processing, loading data into Iceberg for analytics and backing it up in AWS S3.
I designed and maintained a PostgreSQL analytical warehouse using star-schema dimensional modeling, bringing together data from multiple operational sources. Kafka, Debezium CDC, RabbitMQ signaling, and NiFi ingestion flows kept warehouse datasets synchronized for reporting, product analytics, and data science.
At Continental Automotive R&D, I developed embedded automotive software in C under real-time constraints and built Python and CAPL test-automation tools that cut test-campaign preparation and execution from approximately two weeks to under three hours. I also developed processing pipelines for ECU, HIL, and CANoe test outputs, and taught programming and data structures as a teaching assistant at the Faculty of Technical Sciences.
