Staff Machine Learning Engineer, Ads Forecasting & Scenarios

Description

The company runs a community platform built around shared interests and open conversation, with about 100,000 active communities and roughly 130 million daily active unique visitors.

This is a hands-on technical leadership role that owns modeling direction, evaluation standards, and reusable production capabilities for advertiser planning and recommendations.

The goal is to help advertisers see what their campaigns could achieve and make better investment decisions. You will build models that estimate how outcomes shift when budgets, bids, audiences, creative quality, or objectives change, linking supply and demand forecasts, audience availability, response prediction, and marketplace simulation so advertisers can compare options and spot opportunities they would otherwise miss.

Responsibilities

• Turn advertiser questions into models of delivery, reach, performance, and investment tradeoffs, then set technical direction and steer the team's execution.
• Combine forecasts of ad opportunities, audience activity, competition in the marketplace, and predicted response to estimate what delivery and performance are achievable.
• Build models of outcomes under different budgets, bids, audiences, and optimization goals, including how those choices interact with auction behavior and campaign limits.
• Use creative-performance estimates from partner teams to judge how stronger creative could have improved results, and convert that into gains in delivery, efficiency, or advertiser value.
• Apply probabilistic forecasting, multitask learning, transfer learning, or hierarchical modeling to cope with sparse segments, cold start, delayed feedback, and shifting conditions.
• Assess accuracy, calibration, uncertainty, and performance across cohorts. Define which scenarios are supported and when more evidence or experiments are required.
• Compare recorded predictions against mature outcomes, find systematic errors, and check scenario estimates through experiments and real results.
• Work with engineers on reliable training, serving, and monitoring, lead design reviews, mentor engineers, and coordinate with Product, Data Science, Ads Delivery, Audience, and Creative teams.

Requirements

• A record of building and shipping predictive ML systems in advertising, marketplaces, recommendations, search, demand forecasting, or comparable complex fields.
• Solid grounding in statistical learning, probability, and optimization, with experience on large-scale behavioral or transactional data.
• Depth in at least one of: probabilistic or hierarchical forecasting, calibrated conversion or value prediction, multitask or transfer learning, delayed-feedback or survival modeling, or response modeling.
• Sound evaluation judgment, covering temporal validation, calibration, segment-level results, and the assumptions behind estimates under changed conditions.
• Experience with sparse data, cold start, seasonality, distribution shift, selection bias, or partially observed labels.
• Strong programming and software engineering skills, including building and running production ML systems.
• A history of leading unclear technical work, making practical tradeoffs, and growing other engineers.
• Preferred: depth in one or more of the following areas.
• Demand, supply, and inventory forecasting for advertising inventory, marketplace demand, traffic, retail demand, or capacity. Useful methods include probabilistic time series, hierarchical or global forecasting, external-signal modeling, and multi-horizon prediction.
• Auction and investment modeling, such as auction simulation or replay, conditional win-rate and price models, bid or budget response curves, and constrained optimization.
• Scenario evaluation, such as counterfactual or off-policy evaluation, sensitivity analysis, uncertainty propagation, and combining several predictive models.
• Reach and audience estimation, such as population and overlap estimation, probabilistic reach and frequency models, sampling correction, and saturation modeling.
• Advanced predictive methods, such as mixture-of-experts, distributional or quantile prediction, temporal sequence models, and state-space models.

Conditions

This posting may cover more than one career level.

The role is eligible for restricted stock units in addition to base salary and, depending on the position, may also be eligible for commission. For U.S.-based employees the company offers medical, dental, and vision insurance, a 401(k) with employer match, generous vacation time, and parental leave. More details are on the careers page.

Base salary ranges are shared for all U.S.-based postings in every state. Ranges are set by function, level, and country and benchmarked against similar growth-stage companies. Final offers depend on factors such as skills, depth of experience, and relevant licenses or credentials, and may fall outside the range below.

The base salary range is $230,000 to $322,000 USD.

In some roles and locations, interviews are recorded, transcribed, and summarized by AI, and you can opt out before any scheduled interview.

The company is an equal opportunity employer and provides reasonable accommodations for qualified individuals with disabilities and disabled veterans during the application process. If you need an accommodation during interviews, tell your recruiter.

Benefits

• Comprehensive healthcare and income replacement programs.
• 401(k) with employer match.
• Global benefit programs covering workspace, professional development, and caregiving support.
• Family planning support.
• Gender-affirming care.
• Mental health and coaching benefits.
• Flexible vacation and paid volunteer time off.
• Generous paid parental leave.