Enterprise case study · BridgeNext · HR automation system
HR appraisal automation
A serverless pipeline that turns an HR spreadsheet into personalised appraisal letters — rendered as PDFs, stored, and emailed in batches.
- Role
- Software Development Engineer · backend & architecture
- When
- Since 2024
- Links
- Confidential client work
- Excel upload
- Azure Function
- Velocity → PDF
- MSSQL + email
The problem
Appraisal letters were produced from spreadsheet data — repetitive work that grows with headcount and has to be exact for every employee.
What I did
- Contributed to the system architecture
- Built Excel parsing and PDF generation in Azure Functions
- Implemented batch-wise email dispatch and MSSQL persistence
Architecture
- An Azure Function parses the uploaded Excel sheet
- Letter templates live in Azure Blob Storage and are rendered with VelocityJS into PDFs
- Generated letters and their records are stored in MSSQL
- Emails are sent in batches rather than all at once
Key decisions
Templates as data, not code
Keeping Velocity templates in Blob Storage means letter wording can change without a deployment.
Batch the side effects
Batching keeps sends within email-provider limits and contains a failure to one batch instead of the whole run.
Stack
- Azure Functions
- Node.js
- Azure Blob Storage
- VelocityJS
- MSSQL
Client work under NDA — no code, screenshots or client names. The architecture and my role are described as they were.