Sasha User

I’m a Staff Java Engineer specializing in distributed systems and data platforms, with 12+ years at CARFAX architecting high-throughput, low-latency infrastructure. My work focuses on turning complex data workflows into reliable, measurable services—consistently targeting strict performance envelopes like “10ms P99 latency.”

At CARFAX, I led the multi-phase modernization of the core Metadata Platform, cutting metadata query latency from ~35ms to under 15ms while increasing coverage “5×.” I engineered high-throughput storage for 4B+ metadata records using Parquet on Hadoop and built a high-availability dual-cluster architecture with transparent failover, adaptive load balancing, and cross-cluster replication.

I also re-engineered NLP/AI decoding for automotive service/repair data, scaling coverage from 20M to 4B+ records and enabling deep error analysis with Parquet analytical datasets at each stage. Alongside this, I built a flexible data access layer with Spring Boot REST APIs and Hadoop Parquet datasets so enterprise teams could consume metadata without reverse-engineering upstream JSON behavior.

I’m a strong technical leader and open-source contributor, with experience mentoring teams and driving cross-functional adoption through prototypes, benchmarking, and structured knowledge transfer (30+ hours). I authored open-source “Flexible Object IO” to streamline serialization across JSON, YAML, Parquet, Avro, Kafka, RabbitMQ, MongoDB, SQL, and Apache Arrow.