Research Engineer, Data Infrastructure (Language Modeling)

Cartesia is a company focused on developing advanced AI models that learn and interact like humans. Their team, composed of experts from Stanford AI Lab, is innovating with new model architectures such as State Space Models to enable efficient large-scale foundation models. The company is well-funded by prominent investors and benefits from a network of advisors and industry experts, fostering a cutting-edge research environment.

The role of Research Engineer in Data Infrastructure involves creating and maintaining systems that handle large datasets essential for model pretraining. Daily tasks include building scalable data pipelines, processing massive text datasets, and collaborating with research teams to optimize data quality and composition. The engineer will also experiment with data sources and manage external data partnerships to enhance model training.

Candidates should have practical experience with machine learning data infrastructure, including data pipelines, versioning, and large-scale data loading. Strong engineering skills, familiarity with current tools, and understanding of dataset creation for generative models are required. Knowledge of data processing frameworks like Spark, Ray, or Kubernetes and experience with language model pretraining are considered advantageous.