You'll build the models that decide, in real time, what an ad on Connected TV or mobile is worth. They run inside a bidding system handling millions of requests per second, so small gains in accuracy turn into real revenue. This is a hybrid role in New York City with a global ad tech company that's growing its US machine learning team.
What you'll do: Build and improve conversion and lifetime value (LTV) prediction models for real-time bidding Design systems that combine several estimators to set bid values in real time Solve hard modeling problems: extreme class imbalance, delayed attribution, noisy or incomplete labels, and cold-start users Build targeting and identity features from IP-based signals on CTV inventory Own the research cycle from hypothesis to production, including offline experiments and A/B tests Work with engineering, data infrastructure, and operations teams to ship models to production Help set technical direction alongside applied scientists and engineers in Europe What you bring 7+ years in machine learning 5+ years of Python 5+ years building predictive models 3+ years of SQL and statistical analysis Strong experiment design and A/B testing skillsMS or PhD in ML, Statistics, Applied Math, or Computer Science preferred Nice to have 3+ years in ad tech, especially CTV or real-time bidding Large-scale ML systems in production Deep learning, transformers, or sequence models Recommendation systems Scala or Java The details Hybrid, New York City, NYFull-time, direct hire

Also on the board Same function, level within a rung

Level

Senior

Location

New York, NY

Occupation

Data Scientists

Industry

Custom Computer Programming Services

Posted

yesterday

Apply for this role →