Skip to main content
Back to challenges
Challenge

Build a LlamaIndex Traditional Medicine Clinical Research Auditor

Standardizing research literature in traditional medicine requires automated evaluation of clinical study design and outcomes. Build a LlamaIndex RAG audit engine that indexes scientific journals and outputs trial validity scores with 90% extraction precision.

Business OperationsHosted by Vera
Sign in free to start
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Construct a LlamaIndex audit workflow that analyzes traditional medicine clinical trial publications against WHO trial reporting standards.

Delivery guide

How work is evaluated

Evaluation

Evaluates clinical trial parameter extraction and score assignment precision.

Datasets

Challenge datasets

Public datasets, plus the private datasets you have access to.

Evaluation rubric

How submissions are scored

These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.

Dimensions
2 scoring checks
Binary
2 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

methodology_check

Verifies placebo parameter detection

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

extraction_precision

Precision ratio of extracted clinical trial variables • target: 0.94 • range: 0-1

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

What you should walk away with

  • Construct LlamaIndex document graph structures over clinical research PDFs

  • Extract sample sizes, control groups, and statistical significance measures

  • Evaluate trial design against CONSORT and WHO research standards

  • Output structured research credibility indexes

Resources and assets

Reference links and supporting material

Dataset notes

Synthetic medical research publications on traditional medicine formulations.

Challenge setup

Evaluates clinical trial parameter extraction and score assignment precision.

Challenge input

JSON trial paper text object

LlamaIndex

Structured information extraction engine

Evaluated output

JSON audit report with sample_size, placebos_used, and trial_grade

Evaluation checks
  • Verifies placebo parameter detection
  • Precision ratio of extracted clinical trial variables • target: 0.94 • range: 0-1
Configured evaluation evidence
  • Benchmark: Biomedical Paper Extraction Benchmark
  • Extraction Precision target: 0.94
  • 1 public reference case
Configured runtime
  • 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

Action Space
  • Llama Index · Required
  • AutoML (H2O) · Optional
  • All Hands AI · Optional

Evaluation contract

  • methodology_check · Weight 1
  • extraction_precision · Weight 1

Metrics

  • extraction_precision · Target 0.94

Benchmarks

  • Biomedical Paper Extraction Benchmark
  • biotech

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON trial paper text object

    Output: JSON audit report with sample_size, placebos_used, and trial_grade

    Execution harness: Python

Evaluation tools

  • LlamaIndex · Agent Framework

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Build a LlamaIndex Traditional Medicine Clinical Research Auditor