Brian Sanders

About

Backend/ML platform engineer from Snapsheet who owned production Python services and data pipelines for insurance claims, including an AI document classification/summarization FastAPI service on ECS/Fargate processing 1M+ documents/year. Strong in AWS infrastructure (Terraform, CI/CD, secrets/IAM, autoscaling), Glue/PySpark ETL with schema evolution controls, and legacy SAS-to-microservices modernization with safe, feature-flagged rollouts and measurable performance wins.

Experience

Education

Key Strengths

• Built and owned high-throughput FastAPI ML service on AWS ECS/Fargate supporting 1M+ documents/year
• Production reliability engineering (JWT/RBAC, strict timeouts, idempotent retries, clean 4xx/5xx mapping, structured logging with request IDs)
• Infrastructure-as-code and CI/CD ownership (Terraform + GitHub Actions; rolling/blue-green deployments)
• Secure, consistent configuration/secrets management across environments (Secrets Manager/Parameter Store + IAM roles; startup config validation/fail-fast)
• Designed and owned AWS Glue ETL ingesting from ~12 partners with explicit schema versioning and quarantine-based data quality controls
• Handled breaking schema changes without downstream disruption via dual-write/backward compatibility and communicated migration path
• Modernized legacy SAS claims processing into Python microservices with parity testing, stakeholder sign-off, feature-flagged incremental rollout and rollback
• SQL performance optimization with measured impact (65s to <3s) using indexes, query rewrites, partitioning, and batching
• Incident ownership for batch pipelines with mitigation and prevention (partial reruns, increased parallelism, automated retries; reduced similar failures by 80%+)