James Nesbit

Staff Data Scientist Attentive

[email protected] · +1 (646) 823-0740 · Curriculum Vitae

I'm James Nesbit, a Staff Data Scientist at Attentive with a Ph.D. in Economics from NYU (New York University). My work spans experimentation and causal inference, contextual bandits, and production ML systems that turn statistical rigor into measurable business value. Before Attentive, I spent four years at Amazon building forecasting and supply-chain optimization systems. My research background is in econometric theory, matching and optimal transport, and machine learning.

Professional Experience

Attentive

Staff Data Scientist (January 2026–Present) · Senior Data Scientist (June 2025–January 2026)

I lead science across Attentive's brand on-site, experimentation, and conversational AI pillars. I shipped the company's brand-facing experimentation platform — now live across 1,600+ brands and ~134K experiments per year, with company-wide holdouts — designing its analysis stack around variance reduction (CUPED) and always-valid inference to shorten experiments and reduce revenue lost on suboptimal variants. I built AI Grow, a contextual bandit for sign-up popup optimization, and launched Attentive's first customer-facing GenAI agent, which lets customers ask plain-English questions about their performance. I also defined the sign-up popup health and performance metrics that anchor the 2026 on-site strategy.

Amazon

World Wide Return and ReCommerce Planning · Senior Applied Scientist (January 2025–June 2025)

I architected forecasting and supply-chain planning platforms for Amazon's reverse logistics, identifying a $230M planning-accuracy entitlement. I developed interpretable time-series models with built-in decomposition, established a statistical forecast goal-planning framework for Finance, and built causal-inference tooling that quantified the downstream impact of returns-policy changes.

F3 Distribution Optimization and Grocery Innovation (March 2021–January 2025)

I designed and implemented optimization algorithms for transportation planning, replenishment, inventory, and store-layout optimization across thousands of Amazon Fresh and Whole Foods locations, delivering over $100M in cost savings. I built solutions in Java, Python/SQL, and React on AWS infrastructure, and collaborated with forecasting science teams to integrate time-series models into optimization frameworks.

Technical Skills

Publications and Forthcoming

Working Papers

Education

Ph.D. in Economics, New York University (2015–2021)
Advisors: Tim Christensen, José Luis Montiel Olea, Alfred Galichon