MD_ASSIST submitted to OpenAI Build Week
The first working MD_ASSIST prototype is now complete and submitted!
This build demonstrates an end-to-end workflow that transforms fragmented, synthetic patient records into concise, source-grounded clinical context for pre-visit preparation.
Included in this release
- A working clinician-facing prototype
- Synthetic patients and medical records—no real patient data
- Source-grounded responses with visible evidence references
- Safety boundaries separating information preparation from medical judgment
- Documented evaluations for repeatable testing
- A complete public demonstration video
MD_ASSIST was developed as a solo project using Codex and GPT-5.6. Next, I’ll expand evaluation coverage, improve source traceability, and seek feedback from qualified healthcare professionals.
Thank you for following the journey—constructive feedback is welcome.
Log in or sign up for Devpost to join the conversation.