University team project · Work in progress

Multimodal Survey STEM module preview

This is ongoing academic team work, not a solo portfolio product. My contribution focused on the Flask REST endpoints, model-readiness and error handling, session metadata, and integration between the participant module and the analysis pipeline. The public preview shows only the participant-facing STEM module.

The model readiness check runs before the survey. The collector stores questionnaire answers, derived emotion summaries, and analysis metadata; raw webcam video is not stored in D1.

Preview of the Multimodal Survey STEM questionnaire interface

Module preview

Hugging Face Static Space

Collection layer

Cloudflare Worker + D1

Model runtime

Browser ONNX inference