ehsan faraji

Worked on Apache Flink, Paimon, and served as PPMC Member in Fluss (Streaming Data Lakehouse). Architected and maintained high-throughput streaming projects using Kafka messaging protocol and Apache Flink/Spark Structured Streaming for real-time ingestion.

Implemented continuous model retraining workflows using streaming data lake house to support 100+ ML models with real-time feature serving. Developed and maintained real-time feature stores using Apache Flink, Fluss, and Arrow (SIMD) Execution Format for machine learning pipelines.

Managed a team working on the first official release of Apache Fluss on Ververica Cloud as a stream Lakehouse, reaching 10 gigabits and removing more than half of Deduplication ETL. Engineered features and built ML models using Python libraries and Flink CEP for real-time insight and processing.

Developed and maintained scalable ETL/ELT pipelines and collaborated with ML and data analysis teams to design features for fraud detection. Implemented real-time and batch data processing workflows using Kafka and Apache Spark, optimizing over 10 ETL pipelines processing over 10TB of data weekly.

Worked with cross-functional teams to design data solutions for efficient data ingestion, transformation, and storage. Built and optimized OLAP systems with ClickHouse and SQL Server, and automated data workflows using Apache Airflow, reducing data processing time by 30%.

Developed and maintained scalable APIs for clients and providing services to inquiring organizations. Designed, implemented, and maintained Java applications that are typically high-volume and low-latency for mission-critical systems.

Bachelor of Science in Computer Science from Khajeh Nasir Toosi University of Technology, completed from 2017 to 2023.