Senior Researcher

Role overview

Develop and improve foundational speech models used in voice-enabled products. The work spans research, data quality, model evaluation, training performance, and the path from research prototypes to reliable production systems, with close collaboration across research and engineering.

Responsibilities - Research and test methods for improving speech and multimodal language models. - Investigate training data quality, build tools to reveal issues, and measure resulting accuracy gains. - Evaluate production model performance and create testing tools to identify improvements with meaningful user impact. - Work with research and engineering colleagues to productionize prototypes and improve model architectures and inference systems. - Find and resolve performance bottlenecks in model training.

Requirements - Practical experience training or adapting advanced speech or multimodal language models. - Strong Python skills. - Careful analytical judgment: validate results, investigate unexpected changes, and understand the evidence behind conclusions. - Ability to follow problems across data, evaluation, research, infrastructure, and production engineering. - Initiative in identifying high-impact problems and working through them independently. - Interest in building useful analysis and data-processing tools and sharing them with teammates.