MLOps Engineer

Job Requirements

• Design and implement end-to-end computer vision ML training pipelines on AWS SageMaker for model training, validation, deployment, and monitoring
• Architect and build a scalable image data lake solution enabling multi-modal search capabilities (structured metadata, image-to-image, text-to-image) along with data upload capability from edge devices
• Develop vector embedding pipelines for visual content using AWS services and deep learning frameworks
• Create APIs for seamless integration with third-party annotation services and automated dataset creation
• Implement active learning pipelines that intelligently select high-value images for annotation, optimizing annotation ROI
• Build data quality and validation frameworks to ensure consistency across the annotation lifecycle
• Develop infrastructure automation using AWS CloudFormation/CDK for scalable deep learning workflows
• Establish monitoring systems for data drift, annotation quality, and model performance
• Create skeleton frameworks and templates enabling customers to deploy their own deep learning models
• Optimize storage and retrieval mechanisms for large-scale image repositories

Work Experience

• Bachelors or Masters degree in Computer Science, Engineering, or related field
• 5+ years of experience in MLOps or ML Engineering with focus on computer vision applications
• Experience building data lakes or large-scale data repositories for unstructured data
• Strong understanding of vector databases, embedding models, and similarity search algorithms
• Hands-on experience with AWS services (S3, SageMaker, Lambda, Step Functions, Glue)
• Proficiency in Python and experience with PyTorch or TensorFlow
• Experience implementing active learning systems for optimizing annotation workflows
• Knowledge of RESTful API design and integration with third-party services
• Familiarity with annotation tools and workflows for computer vision datasets
• Experience with containerization (Docker) and orchestration (Kubernetes/EKS)
• Understanding of data governance and security best practices for sensitive image data