Axiom — NorthAxis
Axiom is a next-generation, high-performance data observability and log analytics platform that ingests, stores, and queries massive volumes of enterprise event data in real time. Designed as a disruptor to legacy, cost-prohibitive logging solutions, it provides serverless log management that eliminates the trade-off between data retention and operational cost. From the ingestion pipeline to the frontend data table, I built the full product end-to-end.
Modern enterprises generate terabytes of log data daily, but legacy observability platforms price by data volume — forcing teams to either filter out critical telemetry or face unpredictable bills. On the engineering side, the core hurdles were mitigating ingestion bottlenecks during traffic spikes without inflating cloud costs, and executing complex aggregations over billions of rows in under a second without memory-heavy indexing that causes OOM crashes.
Built a stateless ingestion layer using Go with distributed message queues, dynamic batching, and stream compression before committing data to object storage. Designed a custom columnar storage engine that minimizes disk I/O by executing queries directly on compressed data blocks using SIMD instructions and parallel map-reduce workers. On the frontend, built a reactive virtualized data table using React and Web Workers to render millions of log lines smoothly without blocking the main thread.
