Skip to main content
Loading navigation
Back to challenges

LlamaIndex & Galileo ONDC Merchant Inventory Risk Agent

Mitigate regional inventory stockout risks for quick-commerce catalog integrations across Open Network for Digital Commerce (ONDC). You will construct an agentic index that evaluates supplier health, demand signals, and supply disruption risks to trigger automated reorder workflows.

Business OperationsHosted by Vera
Sign in free to start

What you are building

Build a LlamaIndex agentic workflow using Galileo hallucination monitoring to analyze ONDC merchant inventory feeds and dynamically optimize fulfillment routing.

How work is evaluated

Evaluation

Evaluates the agent's ability to process catalog updates and generate accurate, un-hallucinated reorder actions.

Datasets

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

How submissions are scored

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

Schema Validation Test

Ensures output conforms to reorder action JSON schema

Binary check

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

Dimension 2

Reorder Precision

Accuracy of trigger thresholds based on stockout math • target: 0.95 • range: 0.85-1

Binary check

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

Learning goals

  • Implement LlamaIndex agentic routing for dynamic multi-catalog ingestion

  • Integrate Galileo AI telemetry for monitoring hallucination rates in inventory recommendations

  • Construct a consent-aware supplier risk assessment pipeline for ONDC protocols

  • Automate reorder threshold triggers based on predicted regional demand spikes

Resources

Dataset notes

Simulated ONDC grocery catalog feed with fluctuating stock levels and delivery lead times across multiple regional hubs.

Challenge setup

Evaluates the agent's ability to process catalog updates and generate accurate, un-hallucinated reorder actions.

Challenge input

JSON containing store inventory feeds and supplier lead times

LlamaIndex

Provides data indexing and agentic routing tools.

Galileo

Monitors hallucination and agent tool accuracy.

Evaluated output

JSON array of reorder actions with confidence and urgency flags

Evaluation checks
  • Ensures output conforms to reorder action JSON schema
  • Accuracy of trigger thresholds based on stockout math • target: 0.95 • range: 0.85-1
Configured evaluation evidence
  • Benchmark: SupplyChainAgent-Bench
  • Reorder Precision target: 0.95
  • 1 public reference case
Configured runtime
  • Python execution harness
  • Python sandbox (unavailable on Versalist)

Shows the first task, up to three tools, and up to three evidence items.

View technical recipe

Configured tools

Action Space
  • Llama Index · Optional
  • Hamming · Optional
Observation
  • Galileo · Required
Reward / Eval
  • Galileo · Required

Evaluation contract

  • Schema Validation Test · Weight 1
  • Reorder Precision · Weight 1

Metrics

  • Reorder Precision · Target 0.95

Benchmarks

  • SupplyChainAgent-Bench
  • general

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing store inventory feeds and supplier lead times

    Output: JSON array of reorder actions with confidence and urgency flags

    Execution harness: Python

Evaluation tools

  • LlamaIndex · Agent Framework
  • Galileo · Evaluation & Observability

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about LlamaIndex & Galileo ONDC Merchant Inventory Risk Agent