At Replika, I architected production AI data infrastructure using Python, SQL, Apache Spark, and Databricks patterns. I cut end-to-end processing latency from 42 minutes to 11 minutes by optimizing distributed pipelines.
I also designed data services supporting 3.2 million user profiles and established cloud-native architecture principles across AWS, Azure, and GCP. Automated CI/CD validation and deployment controls helped reduce pipeline-related operational incidents by 37%.
At Supersimple, I designed an enterprise analytics platform connecting machine learning services, structured data processing, and reporting. Its ingestion and transformation pipelines reduced average processing time from 28 minutes to 9 minutes.
Earlier, at Ready Player Me and the Estonian Defense Forces, I developed AI application services and secure enterprise software. Across these roles, I worked on data processing, production machine learning integrations, and scalable backend services.
