Ömer Faruk Koç

I’m an AI/ML Platform Engineer focused on building and operating production ML, data, and Generative AI systems that teams can trust. Over 3+ years in financial services, I combine software engineering, data engineering, and MLOps to deliver scalable AI platforms.

I owned and evolved a centralized feature store and dimensional data models that support reusable training, inference, and analytical datasets. I engineered batch and real-time ML data pipelines processing approximately 9M records per day and deployed 10+ production ML models through containerized services.

My work centers on GPU-enabled ML/LLM serving and reliable AI workflows, including RAG applications and Generative AI pipelines (document ingestion, retrieval, context engineering, structured generation, fine-tuning, and multi-agent tool orchestration). I expose inference capabilities through FastAPI services and modernize infrastructure using MLRun, Docker, and Kubernetes.

I also strengthen production reliability with model performance monitoring, data-quality controls, drift detection, inference monitoring, and scheduled retraining. From modular dbt pipelines reducing Oracle ETL runtime from 120 minutes to 30 minutes, to Airflow and CI/CD integrations for delivery, I’m built for turning AI prototypes into maintainable, monitored platforms.