Context
An admission recommendation workflow needs to translate user-entered criteria into a request the backend can process, then present the returned results clearly. Different input requirements make validation and payload construction important parts of the application.
My contribution
I developed Node.js and Express APIs for criteria-based admission recommendations using controllers, services, and repositories.
I built React and Next.js prediction interfaces with dynamic forms, input validation, debounced rank conversion, and API-driven result processing.
Workflow at a glance
- Candidate inputs
- Validation & mapping
- Recommendation API
- Result presentation
Engineering focus
Form-to-API mapping
Form presentation and request construction are separate concerns. Schema-driven fields and payload mapping help connect editable inputs to the format an API expects.
Input feedback
Required-field and numeric checks provide feedback before a request is submitted. Debouncing conversion requests limits repeated calls while a user is typing.
Results and failures
The reviewed interface handles individual request failures separately, allowing successful results to remain visible alongside errors.
Tradeoffs & considerations
- Layered request handling separates HTTP concerns from domain and retrieval logic, while some operations need only a thin service layer.
- Schema-driven inputs reduce repeated form logic, but domain-specific validation still needs explicit handling.
- Debouncing reduces repeated conversion calls; it does not by itself resolve stale responses or prove a measured performance improvement.