Overview
What this challenge is about.
You receive: 100 internal benchmark questions with reference answers; a 50,000-document anonymized RAG index; a curated drug-target-disease KG (~80,000 triples) loaded into a triple store. Design a retrieval pipeline that (1) runs the existing dense retriever, (2) extracts candidate drug/target/disease entities from the question, (3) queries the KG for grounded relationships, (4) feeds both the retrieved passages and the KG facts into the answer generator. Evaluate hallucination rate with a rule-based checker against the KG plus a 20-question human spot-check. Deliver: pipeline code, benchmark notebook, and a 4-page memo.
The Brief
What you'll do, and what you'll demonstrate.
Cut RAG hallucination rate on drug-target-disease relationships by grounding answers in a curated knowledge graph.
Earning criteria — what you'll demonstrate
- Wire a knowledge graph into a RAG retrieval pipeline
- Extract entities from natural-language questions for KG lookup
- Evaluate hallucinations against a structured ground-truth source
- Reason about the limits of KG grounding when the KG is incomplete
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.
AI Engineer
Wiring a KG into a RAG pipeline with measurable hallucination reduction is exactly the work AI engineers do at any high-stakes RAG product.
This challenge sharpens
- kg-grounded-rag
- retrieval-augmented-generation
- entity-linking
AI Solutions Architect
Designing the entity-linking + KG-query layer plus the limits memo is the AI solutions architect's output in regulated-industry RAG.
This challenge sharpens
- kg-grounded-rag
- knowledge-graphs
- retrieval-augmented-generation
AI Safety Researcher
Measuring hallucinations against a structured source is a methodological contribution that safety researchers ship in regulated AI deployments.
This challenge sharpens
- hallucination-evaluation
- kg-grounded-rag
- knowledge-graphs