Main Tasks & Responsibilities

• Update training and test model databases with new or amended synthetic textual and image data.
• Modify and refine machine learning data creation, annotation, and rating guidelines.
• Initiate model training processes using internal tools and command-line interfaces.
• Evaluate the performance of trained models to gauge their efficacy and readiness for deployment.
• Design and develop test and training datasets as per the criteria provided by the project manager and other full-time employees.
• Handle data efficiently, ensuring its integrity throughout the workflow.
• Engage in data relevance tasks, ensuring data sets are aligned with project goals.
• Annotate data accurately, ensuring it adheres to set guidelines.
• Conduct manual quality analysis of model results.
• Recognize error patterns and report anomalies for further investigation.
• Deliver detailed reports on findings, including aspects such as utterance quality, LLM evaluation, ASR bug tracking, and customer pain points to be reviewed by the User Experience Research team.
• Implement basic quality control measures and ensure the reliability of processed data.
• Utilize intermediate data analysis techniques to extract insights and inform decision-making.
• Arbitrate discrepancies effectively, ensuring consistent data quality.
• Apply basic knowledge of natural language processing and linguistics to data processing tasks.
• Ensure linguistic accuracy in all processed and annotated data.

Requirements

• Bachelor's degree in Computer Science, Data Science, Linguistics or Computational Linguistics or a related field.
• Ability to work in a fast-paced, collaborative environment.
• Excellent communication skills.
• Familiarity with command-line tools and interfaces.
• Strong analytical skills with the ability to identify patterns and anomalies.

Additional Information

• This role primarily focuses on English US data sets; however, familiarity with translation or multi-lingual data sets can be a plus for future projects.

Job Data by YubHub.