Audience framework incorporating lookalike modeling and clustering for a national furniture retailer — 43% YoY ROAS improvement, 5M+ incremental reach, and 3–4× Revenue Per Impression improvement.

Fuzzy matching algorithms consolidating millions of parts across supply chain datasets into a canonical supplier catalogue — $7M in procurement savings and eliminated years of reporting inaccuracy.

Built NLP algorithms to screen 100,000+ vendor contracts at Intel — flagging clause-level risk across every supplier. $33M in annual savings. 2020 Gartner High-Tech Chainnovator Prize.