Skip to main content
Back to challenges
Challenge

LlamaIndex Purdue Model OT/IT RAG Pipeline with Together AI

Integrating Level 0-3 operational telemetry with cloud data lakes risks breaking network segmentation protocols. Build a secure RAG metadata indexing agent using LlamaIndex and Together AI to parse Purdue Model architectures and validate data flow compliance. The system must index 1,000 node schemas and flag boundary violations in under 3 seconds.

EngineeringHosted 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 an OT/IT architecture indexing engine using LlamaIndex and Together AI inference to query enterprise Purdue Model metadata without breaching firewalls.

Delivery guide

How work is evaluated

Evaluation

Evaluates RAG retrieval precision, compliance violation detection, and query execution latency.

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

boundary_detection

Verifies illegal data flows crossing DMZ directly are rejected

Binary check

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

Dimension 2

search_time_ms

Query execution time in milliseconds • target: 500 • range: 0-3000

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

  • Implement a LlamaIndex vector store for Purdue Model Level 0 to Level 4 system schemas

  • Configure Together AI inference endpoint for ultra-fast RAG embedding and generation

  • Enforce OT network safety policy checks on query pathways across DMZ boundaries

  • Benchmark retrieval precision and context grounding for industrial asset queries

Resources and assets

Reference links and supporting material

Dataset notes

Purdue Model architecture layout schema defining 1,000 factory asset nodes, assigned Purdue levels, and network connectivity permissions.

Challenge setup

Evaluates RAG retrieval precision, compliance violation detection, and query execution latency.

Challenge input

JSON query specifying source_level, target_level, and proposed protocol

LlamaIndex

Assigned framework for data indexing and enterprise RAG.

Together AI

Assigned inference provider for high-throughput LLM and embedding calls.

Evaluated output

JSON with allowed (bool), boundary_violations (list), context_retrieved (string), search_time_ms (int)

Evaluation checks
  • Verifies illegal data flows crossing DMZ directly are rejected
  • Query execution time in milliseconds • target: 500 • range: 0-3000
Configured evaluation evidence
  • Benchmark: Industrial RAG Security Benchmark
  • Search Time Ms target: 500
  • 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
  • LlamaIndex · Optional
Policy Serving
  • Together AI · Optional
Training Infra
  • Together AI · Optional

Evaluation contract

  • boundary_detection · Weight 1
  • search_time_ms · Weight 1

Metrics

  • search_time_ms · Target 500

Benchmarks

  • Industrial RAG Security Benchmark
  • security

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON query specifying source_level, target_level, and proposed protocol

    Output: JSON with allowed (bool), boundary_violations (list), context_retrieved (string), search_time_ms (int)

    Execution harness: Python

Evaluation tools

  • LlamaIndex · Data Framework
  • Together AI · Inference Engine

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about LlamaIndex Purdue Model OT/IT RAG Pipeline with Together AI