Location: South San Francisco CA (Hybrid, 3 days/week) (Not remote)
Duration: Long term
JD
Strong understanding of machine learning concepts, algorithms, and best practices.
Proven experience in creating, managing, and deploying ML models using core AWS services such as Amazon SageMaker (for model building, training, and deployment), EC2 (for compute instances), S3 (for data storage), and Lambda (for serverless functions).
Experience with AWS Textract for document data extraction.
Demonstrable experience in designing, developing, and maintaining automated data processing and ML training pipelines using AWS Glue (for ETL) and AWS Step Functions (for workflow orchestration).
Proficiency in ensuring seamless data ingestion, transformation, and storage strategies within the AWS ecosystem.
Experience in optimizing AWS resource usage for cost-effectiveness and efficiency in ML operations.
Experience with Amazon Bedrock for leveraging and managing foundation models in generative AI applications.
Knowledge of database services like Amazon RDS or Amazon DynamoDB for storing metadata, features, or serving model predictions where applicable.
Hands‑on experience with implementing monitoring, logging, and alerting mechanisms using AWS CloudWatch.
Experience with AWS container services like EKS (Elastic Kubernetes Service) or ECS (Elastic Container Service) for managing container orchestration.
Experience in implementing scalable and reliable ML model deployments in a production environment.
Practical experience in implementing, deploying, and optimizing Large Language Models (LLMs) for production use cases.
Ability to monitor LLM performance, fine‑tune parameters, and continuously update/refine models based on new data and performance metrics.
Proven ability to create and experiment with effective prompt engineering strategies to improve LLM performance, accuracy, and relevance.
Proficiency in using Docker to package ML models and applications into containers.
Sorry, the previous response was cut off. I will provide the complete JSON again below.