Skip to main content
Loading navigation
Back to challenges

Salalah AI & Hydrogen Knowledge RAG Assistant using LlamaIndex

Middle East power demand is projected to double by 2050, requiring rapid access to energy transition research. Build an enterprise RAG query engine using LlamaIndex to index technical papers from the Salalah Energy Forum and provide grounded answers on hydrogen integration.

Business OperationsHosted by Vera
Sign in free to start

What you are building

Construct an advanced vector RAG index using LlamaIndex to query power demand forecasts and green hydrogen research from the Salalah Energy Forum.

How work is evaluated

Evaluation

Evaluates accuracy of query answer retrieval and faithfulness against Salalah energy paper ground truth.

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

Target Year Match

Verifies that extracted projection target year is 2050.

Binary check

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

Dimension 2

Retrieval Accuracy

Percentage of accurate facts retrieved. • target: 1 • 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

  • Ingest energy research publications using LlamaIndex data connectors

  • Configure vector store indexing with localized metadata filtering

  • Build a query engine with source citation enforcement for power doubling projections

  • Evaluate retrieval faithfulness and answer relevance scores

Resources

Dataset notes

Text corpus of Salalah Forum energy system papers and Rystad Middle East 2050 power demand reports.

Challenge setup

Evaluates accuracy of query answer retrieval and faithfulness against Salalah energy paper ground truth.

Challenge input

JSON object with user query string.

LlamaIndex

Enterprise RAG indexing and document retrieval

Evaluated output

JSON object with retrieved_answer, cited_year, and projected_demand_growth.

Evaluation checks
  • Verifies that extracted projection target year is 2050.
  • Percentage of accurate facts retrieved. • target: 1 • range: 0-1
Configured evaluation evidence
  • Benchmark: GCC Energy RAG QA Benchmark
  • Retrieval Accuracy target: 1
  • 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 · Required
  • LangWatch · Optional
  • Cartesia · Optional

Evaluation contract

  • Target Year Match · Weight 1
  • Retrieval Accuracy · Weight 1

Metrics

  • Retrieval Accuracy · Target 1

Benchmarks

  • GCC Energy RAG QA Benchmark
  • energy

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON object with user query string.

    Output: JSON object with retrieved_answer, cited_year, and projected_demand_growth.

    Execution harness: Python

Evaluation tools

  • LlamaIndex · Data Framework

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Salalah AI & Hydrogen Knowledge RAG Assistant using LlamaIndex