At FAM, I designed and deployed a low-latency CDC data platform using OLake, ClickHouse, Apache Iceberg, and S3, enabling Bronze/Silver/Gold data modeling with 15-minute data availability. I also led the Kubernetes-based platform setup for ingestion and processing.
I built ingestion and analytics pipelines processing 6TB+ of Cloudflare logs, and developed a DuckDB analytics serving layer on Kubernetes (EKS). My work also included Kafka-to-Iceberg ingestion and an automated Spark compaction pipeline on AWS EMR.
At Amazon, I automated backend workflows using AWS services, cutting manual workload from multiple days to under three hours. I also developed full-stack features with Java, React, and DynamoDB, and built a recipe-generation web application that uses food images to identify ingredients and generate recipes.

