We're looking for a Model Behavior Architect to help build Perplexity's AI products and evaluations. You'll sit within our AI team and collaborate closely with research and product teams, designing prompt and context engineering strategies to deliver high quality user experiences across multiple domains and models.
What you'll do
- Context Engineering: Design, test, and optimize context strategies and system prompts that shape answer engine behavior across products, features, and use cases.
- Evaluation Systems: Build automated and semi-automated evaluation pipelines that measure model quality, catch regressions, and scale across product surfaces.
- Model Launch Support: Partner with research and engineering to validate model behavior before and during rollouts, ensuring smooth transitions with no degradation.
- Research & Analysis: Identify inconsistencies and failure modes in model outputs through well-designed research projects — for both internal and production-facing systems.
- Cross-functional Collaboration: Work closely with design, product, and research teams to translate product goals into concrete model behavior requirements.
- Knowledge Sharing: Help engineers across teams build intuition for prompt design, context engineering, and evaluation best practices.
- Staying Current: Track the latest alignment, evaluation, and prompting techniques from industry and academia, and bring the best ideas back to the team.
What you need
- Experience designing evaluations, benchmarks, or metrics for AI systems.
- Strong written and verbal communication skills, particularly in explaining complex concepts to diverse stakeholders.
- Ability to manage multiple concurrent projects in a fast-moving environment.
- Strong experience with Perplexity or other frontier AI models in production settings.
- Demonstrated experience with Python — you'll prototype, debug, automate, and build systems at scale.
- 3+ years of experience working with LLMs in a product or research setting.
