Perplexity is seeking experienced ML engineers to design, build, and optimize the recommendation systems that power core experiences on Perplexity.
To do this, we are reimagining recommendation systems for the LLM era. Our goal is to combine the intelligence of frontier LLMs, the personalization context that comes from real product usage, and the continual learning capabilities of modern recommendation systems. We build systems that draw on past context and connected data sources to deeply understand each user's needs and recommend the actions that help them get the most out of Perplexity.
What you'll do- Own the personalization and ranking behind key product surfaces to make Perplexity more useful and drive impact on core user and business metrics.
- Build user modeling that captures intent, preference, and propensity, and powers more relevant, more personalized experiences.
- Design the decision layer that balances competing objectives to produce the best overall experience for the user.
- Build the data and evaluation foundations that let these systems learn and improve with usage.
- Help shape the technical direction of ranking, recommendations, and personalization at Perplexity.
What we're looking for- Deep, hands-on experience building production recommendation, ranking, or personalization systems at scale.
- Strong ML fundamentals, covering areas such as engagement modeling, model calibration, offline and online metrics, and online experimentation.
- Experience integrating LLMs into ranking, retrieval, or personalization pipelines.
- Taste and judgment for how personalization should work in an LLM-native product, and curiosity about reimagining it from first principles.
- For tech leadership roles, we will also look for prior experience setting technical direction for recommendation/ranking projects.
Nice to have- Experience with large-scale ranking and training infrastructure (multi-stage retrieval and ranking, feature stores, real-time serving).
- Background in user understanding, feed ranking, notifications, growth, or lifecycle modeling.
