Researcher, Post Training

Cartesia is building an AI platform focused on ubiquitous, interactive intelligence running on-device. They are seeking a Researcher to contribute to the Post-Training team, developing methods for preference optimization and alignment of multimodal models. The role involves designing evaluation frameworks, implementing experimental systems, and translating research into production-ready tools.

Requirements

• Deep knowledge of preference optimization and alignment methods, including RLHF and related approaches
• Experience designing evaluations and metrics for generative or multimodal models
• Strong engineering and debugging skills, with experience building or scaling complex ML systems
• Ability to trace and diagnose complex behaviors in model performance across the training and evaluation pipeline

Benefits

• In-person team based in San Francisco
• Fast shipping and novel work