Work Recap 2026-08

Context

The project is a web service which manages clinical trial documents, known as an eTMF(electronic Trial Master File)

The project was planned by an other department. And was developing with outsourcing. On the first day of August, it was decided that the project would be handed over to my team.

I'm was back-end developer, preparing to take over the project. I received a snapshot of backend and frontend source code and started analysis them.

Faced Problems and What I did

The backend was a AI-generated project typically.

Unnecessary AI-generated Markdown files were there and not managed. It seemed would be drifted as the project would be grown.

Test codes were not in the source. And nothing qualified code behaviors. If I needed to edit a class, it might be a mass by the side effect.

The boolean type were configured as a literal character. It had been handled by String type in the project. So the logic tested whether it is true was testing whether it is equal to "Y" as a literal. I would cause confusing to developers as me.

Some God Classes were in there. About 30 classes were excess 1000 lines and about 5 classes were about 3000 lines.

The outsourcing developing was extended to September, The refactoring all codes at that time had risks to integrate the finalized code later. For avoiding the risks, I listed backlogs instead of editing source codes.

Then I focused to build the cloud infrastructure that would be integrate with the app. I had communicated with Infra Manager in my company, specified the architecture and configured it.

I achieved the external infra resources integration to the app with minimal code editing:

• Migration from static API keys based configuration to Delegating Default Credential Providers for S3, RDS, SQS, SES
• Simplified and organized application properties to explicit environments using env post-fixes
• builded a CI/CD pipeline with AWS CodePipeline