Research Staff - Synthetic Data

Role overview Develop research methods and data pipelines that improve training for information retrieval systems. The focus is synthetic data across modalities: designing the process from generation through quality assessment and refinement, then using feedback to address model performance gaps.

Responsibilities - Design synthetic data generation pipelines for different modalities. - Conduct applied research in information retrieval and contribute to retrieval model improvements. - Build feedback loops that continuously improve generated training data. - Identify model performance gaps with engineering and product teams, then develop data to address them. - Refine existing datasets through filtering or modification to improve training efficiency and downstream results. - Share research through papers, technical articles, or conference presentations.

Requirements - The available description points to a research role focused on synthetic data for retrieval, but the source is shortened before its full requirements are shown. No specific degree, years of experience, or programming language requirements can be confirmed from the provided material.

Nice to have - Open-source contributions to data, information retrieval, NLP, or machine learning libraries. - Familiarity with retrieval systems and late interaction methods.

Benefits and work setup - Full-time, remote role. - Compensation includes equity; health, dental, and vision coverage are offered. - Visa sponsorship and relocation support are listed, along with professional development, conference attendance, transportation, wellness, and food support.