At Conviva, I developed the storage and checkpointing layer for a Rust-based streaming platform that processed 200k events/sec (600 Mbps). I also engineered custom operators for a DAG-based timeline engine to support scalable timeseries analytics.
At Walmart Global Tech, I migrated demand-forecasting Hive queries to Spark DataFrames, reducing execution time from 50 hours to 5 hours and hardware costs by 15%. I also designed an item ingestion pipeline using Spark Streaming and Kafka Streams that processed 20 million items in 2 hours, with Apache Ignite integration for real-time caching.
I'm a Principal Software Engineer at Atlassian and contributed to Apache OpenDAL. I share engineering learnings and architectural insights on my personal blog.
