Client Interface
Project phase: Deployment and Maintenance
You build the end-to-end, user-facing front end that lets someone actually interact with your machine learning system: entering input, triggering a prediction through your API, and seeing the result presented in a way a non-technical user understands, including an honest indication of uncertainty or model limitations. Your design choices are shaped by consulting a client or stakeholder rather than invented alone, and you test the interface with at least one representative user and document how their feedback changed it.
Starting Points
Key Points
- End-to-End Functionality: The interface successfully sends data to the API and displays the resulting prediction.
- User-Centric Design: The interface is intuitive for a non-technical user and clearly explains what the output means (including uncertainty).
- Stakeholder Integration: You can demonstrate how feedback from your client or stakeholders directly influenced your design decisions.
- Systematic Validation: You have conducted a user test with at least one representative user and documented the results.
- Iterative Improvement: You have documented the feedback received and implemented specific changes based on that feedback.
- Ethical Awareness: The interface addresses the impact on the user, such as how model limitations are communicated to prevent misuse.
- Documentation: All design decisions, stakeholder interactions, and test results are clearly recorded in a standalone markdown document in GitLab.