LlamaIndex & Claude 4 Opus Senior Housing Operator Intelligence Engine
Senior housing requires complex regulatory, demographic, and operational analysis across massive document stores. Build an advanced RAG retrieval engine using LlamaIndex powered by Claude 4 Opus to query municipal licensing and operational guidelines accurately.
What you are building
The core problem, expected build, and operating context for this challenge.
Develop an enterprise LlamaIndex RAG application with Claude 4 Opus to index and query senior housing regulatory and operational filings.
How work is evaluated
Evaluates factual recall and citation precision for senior housing compliance queries.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
citation_accuracy_test
Checks if provided section citations match actual regulatory code sections.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
faithfulness_score
RAG faithfulness score assessing absence of hallucinated requirements. • target: 0.95 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Construct a LlamaIndex VectorStoreIndex over senior housing state regulations
Configure Anthropic Claude 4 Opus as the primary LLM engine for reasoning
Implement advanced retrieval strategies including sub-question query engines
Evaluate answer faithfulness and retrieval recall on complex regulatory queries
Reference links and supporting material
Corpus of state regulatory handbooks and licensing requirements for senior housing facilities across 10 states.
Challenge setup
Evaluates factual recall and citation precision for senior housing compliance queries.
Challenge input
JSON query specifying jurisdiction and facility class
LlamaIndex
Provides structured document indexing and sub-question query capabilities.
Claude 4 Opus
Delivers superior reasoning over detailed regulatory text.
Evaluated output
JSON response with answer_text, section_citations, and compliance_flag
- Checks if provided section citations match actual regulatory code sections.
- RAG faithfulness score assessing absence of hallucinated requirements. • target: 0.95 • range: 0-1
- Benchmark: HealthcareRegulatoryRAGBench
- Faithfulness Score target: 0.95
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- Llama Index · Required
- ElevenLabs · Optional
- Claude 4 Opus · Optional
Evaluation contract
- citation_accuracy_test · Weight 1
- faithfulness_score · Weight 1
Metrics
- faithfulness_score · Target 0.95
Benchmarks
- HealthcareRegulatoryRAGBench
- general
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON query specifying jurisdiction and facility class
Output: JSON response with answer_text, section_citations, and compliance_flag
Execution harness: Python
Evaluation tools
- LlamaIndex · RAG Framework
- Claude 4 Opus · LLM Model
Recipe freeze marker
No recipe freeze time is recorded.