Manager II, Machine Learning Engineering, Ads Identity Modeling

Pinterest • San Francisco, WA

14d • Hybrid

About The Position

Ads Identity Modeling (AIM) sits within the Ads Conversion Modeling group in the Ads Performance org and focuses on identity and conversion modeling that improves conversion ads performance and unlocks new optimization objectives in close partnership with Ads Measurement and ranking teams. We treat conversion data as first‑class: growing resilient conversion signals and amplifying their usage across Pinterest to drive measurable impact in ads delivery and beyond. The team owns core identity and conversion‑linked ML systems, including User Match Prediction (UMP) , and is also responsible for building models that focus on driving incremental conversions. What you'll do: Lead the roadmap and technical direction for User Match Prediction (UMP) —from features and training data to model deployment—to increase match rate/quality and strengthen downstream attribution and optimization. Define and productionize a model to generate a user match confidence score , and use it to optimize label weighting and feature construction in conversion models (e.g., oCPM/ROAS). Develop and ship models that increase incremental conversions , partnering with Data Science and Product to set success criteria and measurement plans. Partner with Ads Measurement, Ads Signals, Ranking, and ML Infra to translate identity improvements into reliable online impact and strong operational readiness. Guide model architecture evolution (e.g., feature interactions, multi‑task learning, serving freshness) and iterate on UMP initiatives that expand coverage and resilience (CAPI, WAU attribution, IP‑loss mitigation). Contribute to the longer‑term Identity Graph (IDG) collaboration (multi‑hop identifiers, candidate generation and ranking) to improve identity resolution and downstream utility. Hire, mentor and grow ML engineers in ads quality and partner teams and help to uplevel ML talent across the company.

Requirements

• MS or PhD degree in Computer Science, Machine Learning, Statistics or related field.
• 6+ years of industry experience with related background in Ads conversion modeling/products or Ads measurement modeling/science.
• 2+ years of experience as TLM or EM managing an engineering or applied science team.
• Strong software engineering and mathematical skills with knowledge of machine learning.

Responsibilities

• Lead the roadmap and technical direction for User Match Prediction (UMP) —from features and training data to model deployment—to increase match rate/quality and strengthen downstream attribution and optimization.
• Define and productionize a model to generate a user match confidence score , and use it to optimize label weighting and feature construction in conversion models (e.g., oCPM/ROAS).
• Develop and ship models that increase incremental conversions , partnering with Data Science and Product to set success criteria and measurement plans.
• Partner with Ads Measurement, Ads Signals, Ranking, and ML Infra to translate identity improvements into reliable online impact and strong operational readiness.
• Guide model architecture evolution (e.g., feature interactions, multi‑task learning, serving freshness) and iterate on UMP initiatives that expand coverage and resilience (CAPI, WAU attribution, IP‑loss mitigation).
• Contribute to the longer‑term Identity Graph (IDG) collaboration (multi‑hop identifiers, candidate generation and ranking) to improve identity resolution and downstream utility.
• Hire, mentor and grow ML engineers in ads quality and partner teams and help to uplevel ML talent across the company.

What This Job Offers

• Job Type: Full-time
• Career Level: Manager
• Education Level: Ph.D. or professional degree