Machine Learning Engineer

About the role

We’re hiring an ML engineer to own the recommendation engine that decides, in real time, which ad reaches which user at which moment across millions of daily interactions and tens of millions in annualized ad spend. This is a full-stack ML role, you'll go from data pipelines to model architecture to production serving, with direct business impact at every layer.

Responsibilities

• Design and ship a low-latency ad ranking system (retrieval ranking reranking) that selects the optimal campaign and creative for each ad opportunity, balancing advertiser ROAS against user experience
• Architect the data pipelines and feature stores that power continuous model training across reward signals
• Build representations of user behavior from conversational data, engagement history, and contextual signals (geo, device, session context, characters interacted with)
• Build the stack for sub-second latency and cost efficiency, given tight per-impression unit economics

Requirements

• 0-6 years of ML engineering experience. Cracked new grads welcome.
• You've shipped a 1+ ML system in production. Not just research or notebooks.
• Backend depth across data architecture, feature pipelines, and serving infrastructure end to end
• Hybrid infrastructure + ML background
• Zero-defect mindset and meticulous attention to latency, scalability, and reliability
• Comfort with ambiguity. Some interesting open problems (delayed rewards, fatigue modeling, cold start).
• Bias toward shipping. Early-stage pace. Not a 9-to-5.

Qualifications

• In based in SF or willing to relocate quickly. In-person preferred.