I’m a staff machine learning engineer at Superhuman (formerly Grammarly), where I build LLM-powered features for Superhuman Go, our proactive AI assistant. Over the past decade I’ve taken ML systems from prototype to production at Twitter, EvolutionIQ, and IBM Research — spanning LLMs and NLP, recommendation and ranking, and classical methods like gradient-boosted trees and survival analysis — with a high bar for model quality, performance, and Responsible AI. My work spans 50+ published papers, 7,600+ citations, and a dozen patents.