Google · New York

• Sole PM on a 30-engineer team making generative AI trustworthy enough for Google to ship — the grounding & factuality behind AI answers across Search and Gemini, used by billions and built on by 20+ product teams.
• Grounding: corroborates the model's claims with citations — verifiable evidence behind each AI answer — cutting ungrounded responses 50% (30%→15%), in real time (120s → under 20s).
• Factuality: the bar new Gemini models and AI features must clear to ship — our autorater and eval agents (LLM-as-judge, at Google scale) measure accuracy before launch and feed model training, contributing to 16% fewer factuality errors across Search.
• Multimodal: extending grounding to images and video — verifying claims about non-text content, where visual search (Lens) is growing fastest.
• Use agents to run my product area solo — they handle the busywork, so I prototype, ship, and iterate fast, pairing product taste with the technical depth to launch AI products people love and trust.

Publicis Sapient · New York

• Marriott — drove acquisition and conversion on the 0→1 launch of Homes & Villas premium rentals: +288% traffic, +7% conversion via AI personalization; scaled to 140K+ rentals (147M+ members).
• Walgreens — owned loyalty, retention, and revenue: relaunched the program as an AI wellness companion — $1B+ revenue, 105M members, +20 NPS.
• Albertsons — led the Google Cloud partnership (conversational commerce on Search, Maps, Assistant) and an ML customer-data platform across 2,250+ stores.
• U.S. Bank — shipped a top-rated banking app in under 8 months; 5-star, ranked #2 nationally.