•Build LLM application services with streaming batching and backoff
•Build and maintain CI CD pipelines for apps infrastructure and ML pipelines
•Build data ingestion and processing pipelines with Glue and Lake Formation
•Build offline and online eval harnesses for LLM systems
•Design embedding and chunking pipelines for RAG
•Design provision and maintain AWS infrastructure using infrastructure as code
•Enforce safety privacy and compliance with guardrails and access controls
•Implement observability using telemetry and dashboards
•Implement prompt tool and agent execution flows using LangChain
•Manage AWS identity and access management and encryption keys
•Manage IoT deployments including secure messaging and over the air updates
•Operate secure AWS networking with VPC and PrivateLink
•Operate vector search and tune retrieval recall latency and cost
•Optimize retrieval context window caching and inference performance
•Stand up and operate LLM inference endpoints
•401k
•Health insurance
•Paid Holidays
•Paid time off
•Phone stipend
•Wellness stipend
Технологии: AWS, AWS Athena, AWS Bedrock, AWS CDK, AWS CodeBuild, AWS CodePipeline, AWS ECS, AWS EKS, AWS Glue, AWS Lambda, AWS SQS, AWS Secrets, AWS Secrets Manager, AWS Step Functions, Airflow, Amazon Aurora, Amazon Aurora PostgreSQL, Amazon S3, Amazon SageMaker, Aurora PostgreSQL, Autoscaling, Bedrock Knowledge Bases, CI/CD, CloudWatch, Data redaction, Embeddings, EventBridge, Function Calling, GitHub Actions, Guardrails, IAM, Infrastructure as Code, KMS, Knowledge bases, Lake Formation, Langchain, OpenSearch, OpenSearch Serverless, OpenTelemetry, PGVector, PII Detection, Pinecone, PrivateLink, Quantization, RAG, Rate Limiting, Redis, Retrieval-Augmented Generation, Secrets Manager, Step Functions, Terraform, Token budgeting, VPC, VPC Endpoints, Vector Databases, “as-code”