At Cyvers, I own detection and intelligence systems that have helped prevent hundreds of millions of dollars in losses, including a real-time anomaly engine screening tens of millions of blockchain transactions daily. I’ve built malicious-bytecode classifiers, graph-based fraud tracing, scam-site discovery pipelines, and an expert-in-the-loop labeling tool with more than 140,000 manually labeled transactions.
Previously at SKF Group and StackTome, I improved predictive-maintenance AutoML, vibration-analysis automation, recommendation engines, review analysis, and business-metrics prediction systems. I enjoy turning research into production ML products, from model design and evaluation through scalable deployment and mentoring junior data scientists.
