I architect the Large Behavioral Model generating 10B+ daily predictions across 1.8B+ devices for Fortune 500 marketing decisions. Previously led Adobe GenStudio's guideline-aware AI systems; 9 granted patents (15+ filed) in behavioral ML and GenAI.
I architect ZeroToOne's Large Behavioral Model, the platform unifying physical and digital behavioral traces across 1.8B+ devices, generating 10B+ predictions daily for Fortune 500 marketing intelligence.
At Adobe, I led GenStudio's guideline-aware AI systems. Delivered semantic knowledge graph grounding that enabled Adobe's first fully AI-generated marketing campaign and 60% content velocity improvement.
Carnegie Mellon Ph.D. in privacy-preserving behavioral modeling. Published in ISR, DMKD, KDD, Management Science with 300+ citations. Research deployed in ZeroToOne's production systems.
California. Wife. Cat named Binxy.
Ph.D. in Information Systems
2016 - 2021M.Sc. and B.Sc. in Mathematics and Computing
2009 - 2014Head of AI & Platform
Led technical diligence on ZeroToOne's acquisition of GroundTruth and WeatherBug; lead the platform integration.
2025 - PresentSenior Machine Learning Engineer
2021 - 2025Founding ML Engineer
2017 - 2021Research Scientist
2014 - 2016• Multi-modal, multi-task foundational models
• Spatio-temporal trajectory modeling
• Unified behavioral representations
• Intent prediction
• Knowledge graph-augmented generation
• Parameter-efficient fine-tuning
• Fact verification systems
• Prompt optimization
• Distributed data processing
• Automatic feature engineering
• Model monitoring at scale
• Continuous learning systems
Enterprise AI systems delivering measurable business outcomes.
Large Behavioral Model unifying physical and digital signals for predictive marketing intelligence.
Impact: 3x ranking lift • 10B+ daily predictions • 48% lower cost per visit
Semantic knowledge graph grounding for guideline-aware content generation.
Impact: 60% content velocity gain • Adobe's first AI campaign • Brand safety at scale
Differential privacy techniques enabling enterprise mobility intelligence.
Impact: Privacy-preserving analytics • ISR 2023 • COVID-19 response deployment
Subspace characterization algorithms for interpretable anomaly detection.
Impact: Enterprise fraud detection • Human-interpretable rules • DMKD 2018
Research deployed in production. Patents protecting enterprise AI systems.
Interpretable anomaly detection using subspace rules for enterprise fraud and risk analysis • 68 citations
Privacy-preserving behavioral prediction balancing individual privacy with enterprise mobility intelligence • 52 citations
Large-scale behavioral analysis of privacy decisions enabling crisis response systems • 33 citations
Attention architecture for jointly extracting entities and relations from enterprise documents
5 filings • Enterprise content safety and generation
7 filings • Privacy-preserving ML for enterprise
3 filings • Enterprise knowledge representation
Total: 9 granted • 15+ filed across behavioral ML, GenAI, and privacy
Advisory. Speaking. New opportunities.