Alexander Ivaniuk

I have designed architecture, developed and delivered multiple core components of LinkedIn infrastructure, including planet-scale online services, nearline and messaging systems with up to 60 Mln QPS, data systems processing 100+ Trillion records a day. I have led, developed multi-year roadmaps and delivered large projects with 30+ teams in them impacting 1,800+ LinkedIn engineers. I have extensive hands-on experience with the entire LinkedIn’s serving and data infrastructure stack, including online systems, streaming, ingestion, data stores, offline data processing systems, compliance, lineage, analytics and reporting, including open-source systems as Kafka, Pinot, Spark, Venice, gRPC, DataHub, Gobblin and numerous internal systems. I have been involved in building managed (cloud-like) nearline and offline data processing solutions, and recently I became a tech lead of ML Ops at LinkedIn. I continue to maintain coding experience by prototyping new projects and delivering critical, most technically complex parts of software. Key Accomplishments: * Scaled experimentation and reporting platform via redesign, re-architecture and continuous optimizations. Online systems: 100x QPS (65 Mln QPS peak) and 10,000x call volume with 16x less latency reduction and increasing reliability 99.9% -> 99.999%. Reporting systems: 2,000x load at 60x less latency over Trillions of rows of data. * Developed a large scale, managed user analytics collection and attribution processing system in a company-wide effort (1 Mln messages/sec) * Led and delivered experimentation platforms for enterprise products (15 teams involved) * Biggest personal contributions: 1) translator and optimizer from a common LinkedIn SQL into multiple runtimes; 2) LinkedIn Experimentation DSL