About this role
Amazon’s Worldwide Returns, ReCommerce & Sustainability team seeks a Science Annotation Ops Analyst in Hyderabad to build the data pipelines, dashboards and annotation tools that support high-quality training data for machine learning models. The role works with Program Managers, Science Annotation Specialists and other teams across the annotation data lifecycle.
Responsibilities
- Design and maintain pipelines that extract, parse, validate and consolidate annotation metrics from AWS and SageMaker, with attention to data quality and auditability.
- Build and operate production dashboards on EC2 for ingest, validation, scoring and publishing, using tools such as Plotly Dash or Streamlit.
- Implement least-privilege, cross-account data ingestion with IAM, STS, Lambda and API Gateway.
- Develop SageMaker Ground Truth labeling templates in HTML and JavaScript, translating written procedures into validated annotation interfaces.
- Automate monthly consolidation, historical backfills and scheduled jobs; maintain alerting, backups and deployment workflows. Use Python, pandas, boto3 and SQL to process and validate data, and Git and Linux command-line tools for deployments. Serve as the team’s technical point of contact.
Basic qualifications: A bachelor’s degree obtained within the last 12 months in computer science, machine learning, engineering or a related field, or experience with data scripting languages such as SQL, Python or R, or statistical or mathematical software such as R, SAS or Matlab. Experience with AWS services including EC2, Lambda, S3, DynamoDB and SQS, and with building web-based dashboards using common frameworks, is also required.
Preferred qualifications
SageMaker Ground Truth experience, familiarity with machine learning or annotation operations, and experience standardizing metrics and processes across teams.