Pacific Life
Moved Pacific Life's scheduled batch work from Quartz to serverless, and automated claims with OCR.
- Pacific Life Insurance
- Contract Management and MyAccount
- Cloud and solution architect, full-stack developer
- 2 people
What it is
Development, enhancement and predictive analysis for contract and claims management in the MyAccount platform, including the migration of scheduled processing to a serverless batch architecture on AWS.
The problem
Batch processing ran on an in-application Quartz scheduler, which tied jobs to application servers and limited how they scaled. Claims also relied on manual review of scanned documents.
Architecture
Requests flow left to right. Hover a component for its role.
- MyAccount portal client
- Spring Boot 3 services service, on Amazon EKS
- Amazon RDS data
- EventBridge schedules infrastructure, replaced Quartz
- Lambda batch jobs worker
- SNS and SQS queue
- S3 data, scanned claim documents
- OCR and predictive analysis ai, SageMaker, Bedrock
Key decisions
Schedules move out of the application
Quartz jobs became Lambda functions triggered by EventBridge and reporting through SNS, so batch work scales on its own and no longer competes with user traffic.
Audit documents automatically
Scanned claim documents are read with OCR and checked automatically, so reviewers handle exceptions instead of every page.
Highlights
- Led the Quartz to serverless migration on Lambda, EventBridge and SNS.
- Built OCR-based document auditing into the claims process.
- Predictive analysis of contracts and claims with SageMaker and Bedrock.