Software Engineer, Cloud Infrastructure (Multiple Seniority Levels)

Tasks

• Build CI CD and testing pipelines
• Build ingestion and processing pipelines
• Build knowledge bases and data syncs
• Build prompt and tool execution flows
• Create offline and online eval harnesses
• Deploy and operate model endpoints
• Design AWS infrastructure using IaC
• Design RAG chunking and embedding pipelines
• Enforce PII detection redaction and access controls
• Evaluate and optimize inference workflows
• Implement LLM APIs with streaming and batching
• Implement guardrails rate limits and provider routing
• Instrument telemetry with CloudWatch and OpenTelemetry
• Integrate retrieval services into user features
• Manage IoT deployment and OTA updates
• Manage KMS Secrets Manager and audit logging
• Manage data lineage and cataloging with Glue and Lake Formation
• Operate secure networking VPC and IAM
• Operate vector search and tune recall latency and cost
• Optimize storage and bulk data movement in S3 and Glacier
• Orchestrate data workflows with Step Functions or Airflow
• Provision and maintain cloud services
• Troubleshoot cloud and ML services for uptime and performance
• Tune retrieval context window caching and inference performance
• Use Athena for ad hoc analysis

Perks/Benefits

• 401k
• Healthcare coverage
• PTO
• Phone stipend
• Wellness stipend

Skills/Tech-stack

AWS | AWS Bedrock | AWS CDK | AWS CodeBuild | AWS CodePipeline | AWS Glue | Airflow | Amazon SageMaker | Athena | Aurora PostgreSQL | Autoscaling | Batch | CI/CD | CloudFront | CloudTrail | CloudWatch | DynamoDB | ECS | EKS | Event Driven | Event-driven architecture | EventBridge | GPU Computing | GitHub Actions | Glacier | Glue | Guardrails | IAM | Inference Server | KMS | Lake Formation | Lambda | Langchain | OpenSearch Serverless | OpenTelemetry | PGVector | PII Detection | PII Redaction | Pinecone | PostgreSQL | PrivateLink | Quantization | Rate Limiting | Redis | S3 | SQS | Secrets Manager | Step Functions | TensorRT-LLM | Terraform | Triton Inference | Triton Inference Server | VPC | VPC Endpoints