What You'll Do Own the technical direction of the Decisioning & Optimization team: architecture reviews, incident leadership, capacity planning, and scaling Architect and evolve the real-time ad decisioning optimization path: multi-stage auction, ranking, scoring, bidding, and pacing under strict latency and throughput constraints Scale our ads model serving infrastructure to support dozens of concurrent hot-path ML models with sub-20ms P99 inference, including config-driven model routing, multi-model lifecycle management, fallback tiers, and calibration serving Work closely with Science and Platform teams, ensuring seamless model productionization and algorithm deployment Build out various simulation and containerized testing frameworks to enable offline validation of marketplace changes before live rollout Design and implement real-time pacing systems that drive budget delivery accuracy across campaign lifetimes Develop and scale goal-based delivery optimization, enabling dynamic allocation of budget and inventory across multiple demand channels to maximize advertiser outcomes Drive modularization and platform-thinking: build reusable components and clean interfaces that let the team move faster Drive operational excellence: reliability, observability, deployment automation, capacity planning, and incident leadership across the optimization and broader ad serving stack Skills & Experience We're Seeking 10+ years building distributed systems and backend services at large scale; 3+ years in the ads domain Deep experience with ML model serving infrastructure: scaling real-time inference on the hot path at high QPS with sub-20ms P99 latency, including model deployment pipelines, feature hydration, and fallback strategies Built and operated core ad tech systems: ad servers, bidders, pacers, or ranking and scoring components Designed APIs, platform abstractions, and data models that enable seamless interoperability across a multi-team ads platform Strong understanding of ad serving concepts: inventory management, frequency and recency capping, member ad experience quality, and supply-demand dynamics Track record of technical leadership across multiple teams, setting architectural direction and influencing cross-functional roadmaps Comfortable at the intersection of engineering, data science, and product, translating ML research and algorithms into production systems Demonstrated ability to operate in the environment which is a mix of big-tech scale and startup speed, taking projects that normally take years and delivering production-ready results with tight timelines Nice to Haves Experience with auction mechanics: first-price, second-price, reserve pricing, bid shading, and marketplace competition dynamics Multi-stage ranking systems (retrieval, scoring, reranking), podding and ad break planning Built or improved budget pacing and delivery control systems Yield optimization, inventory forecasting, dynamic pricing, fill rate optimization, and demand/supply allocation strategies Familiar with CTV constraints: server-side ad insertion, live event ad serving at scale Experience with experimentation infrastructure: A/B testing, holdout groups, interference-aware marketplace experiments Built simulation or counterfactual testing platforms for marketplace or auction systems * Strong background in resiliency and reliability: ensuring system availability under extreme load (live events, traffic spikes) Generally, our compensation structure consists solely of an annual salary; we do not have bonuses. Our work spans ML model serving infrastructure, ranking and scoring, auction mechanics, budget and pacing systems, and goal-based delivery optimization along with podding, traffic shaping models, and more.