Perplexity is looking for an Applied AI Engineer to design, build, and iterate on cutting-edge agents powering our core experience in Perplexity Computer. Working in this mission critical team, you will develop frontier context layer applications - fulfilling the curiosity of millions of users across the globe.
Key Responsibilities
Apply state-of-the-art ML and LLM techniques to solve problems spanning:
Personalization (LLM memory, context summarization, retrieval and ranking);
Contextual recommendations and Monetization applications
Build frontier agent capabilities on top of Perplexity Computer
Build auto research harness for both offline and online techniques, designing experiments and metrics that provide deep insight into quality and impact.
Own the entire model lifecycle from research to production: data analysis, modeling, evaluation, offline/online A/B testing, and iterative improvement and build autonomous harness for agent squad to explore different problem spaces.
Collaborate cross-functionally with engineers, PMs, data scientists, and designers to ensure our AI drives meaningful product improvements.
Stay at the forefront of ML/AI innovation by evaluating and incorporating emerging research and algorithms into the product lifecycle.
Preferred Qualifications
5+ years experience building and shipping robust AI products for large-scale, user-facing or data-driven products.
Strong software engineering skills (Python, production-quality codebases, collaborative development) and experience using agentic coding tools for large scale parallel developments.
In-depth experience with the full AI lifecycle: data analysis, rigorous evaluation, and ongoing monitoring/improvement.
Proven collaborator and communicator; excels in high-velocity, cross-functional teams.
Curious, driven by end-user/product impact, and passionate about advancing the state of applied ML and AI.
BS, MS, or PhD in Computer Science, Engineering, or related field (or equivalent experience).
Bonus Points For
Experience with LLM context engineering or harness engineering.
Experience in mid-training or post-training frontier open source models
Experience in large scale user-centric and content-centric personalization challenges (user modeling, retrieval, content ranking, etc).