Applied Scientist II

Our team (Signals Modeling) builds the core intelligence that understands and predicts how users interact with ads - from the first impression through clicks, post-click engagement, and downstream business outcomes. We design and train transformer-based models with billions of parameters that power ad ranking, pricing, and optimization across large-scale consumer surfaces. These models go well beyond simple click prediction: they reason over long user histories, rich ad and content representations, and heterogeneous event streams to infer user intent and advertiser value, even when ground truth signals are sparse or partially unobservable. The team owns end-to-end ML systems, including large-scale data and label construction, representation learning, multi-task and proxy objectives, calibration, and rigorous offline and online evaluation. We build sophisticated training pipelines that transform weak signals (e.g., page visits, dwell time, or engagement events) into high-quality learning targets and deploy models that remain robust under delayed conversions and shifting marketplace dynamics. Engineers and scientists on the team work at the intersection of deep learning, large-scale experimentation, and marketplace economics, shipping production-grade models and data pipelines that directly drive revenue and advertiser ROI. This is a hands-on role with real ownership: you’ll help shape next-generation transformer architectures, push the limits of scalable training and serving, and see your models make measurable impact in one of the world’s largest ads ecosystems.