Overview
What this challenge is about.
Combine traditional NLP (section segmentation, sentence parsing) with LLM extraction (small open model + structured-output enforcement). Build the pipeline so every extracted field has a span pointer back to the original document (auditability requirement). Evaluate on a 100-summary held-out set against clinician-curated gold JSON: per-field precision/recall, particularly for medications. Report critical-medication miss rate and dose-extraction accuracy. Write a 5-page clinical-handoff doc covering known failure modes and required human review.
The Brief
What you'll do, and what you'll demonstrate.
Build a clinically reliable structured-extraction pipeline for discharge summaries with auditable span pointers and proven miss-rate reduction.
Earning criteria — what you'll demonstrate
- Combine traditional NLP and LLM-based extraction for reliability
- Build auditable spans linking structured fields to source text
- Evaluate clinical NLP with field-level precision/recall
- Communicate failure modes to a clinical audience
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
Careers
Roles this prepares you for.
Real titles. Real skill bridges. Pick the one closest to your trajectory.
NLP Engineer
Building clinically-reliable structured extraction with auditable spans is the NLP-engineer work that healthtech AI companies need urgently.
This challenge sharpens
- structured-extraction
- clinical-nlp
- parsing
AI Engineer
Combining traditional NLP and LLMs with schema-enforced decoding is the AI-engineer pattern every regulated AI team adopts after the first hallucination incident.
This challenge sharpens
- llm-tool-use
- structured-extraction
- parsing
Applied AI Scientist
Designing audits and miss-rate analyses for clinical extraction is the applied-AI work that determines whether a healthtech AI product can ship.
This challenge sharpens
- clinical-nlp
- evaluation
- structured-extraction