Build a Multi-Site Battery Storage Portfolio Manager with LlamaIndex
Managing performance across three BESS deals worth $867m requires unified multi-site telemetry aggregation. Build an agentic RAG solution using LlamaIndex workflows to track battery degradation, query maintenance manuals, and optimize warranty coverage.
What you are building
Utilize LlamaIndex Workflows and agentic tools to monitor multi-site BESS telemetry and match fault logs against warranty terms.
How work is evaluated
Evaluates fault code diagnosis accuracy and warranty claim generation correctness.
Datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
warranty_coverage_check
Verify warranty status is correctly identified as covered
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
diagnosis_accuracy
Accuracy of fault code classification against manuals • target: 0.98 • range: 0.9-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Learning goals
Construct LlamaIndex Event-driven Workflows for multi-site telemetry ingestion
Index technical warranty manuals and cell degradation lookup tables
Automate fault diagnosis when battery cell voltage variance exceeds tolerance limits
Generate automated warranty claim submissions for regional BESS suppliers
Resources
SCADA fault logs, battery cell SOH degradation curves, and manufacturer warranty contracts across 3 Saudi BESS sites.
Challenge setup
Evaluates fault code diagnosis accuracy and warranty claim generation correctness.
Challenge input
JSON BMS telemetry log
LlamaIndex
Offers event-driven Workflow architecture for multi-document RAG
Evaluated output
JSON warranty assessment
- Verify warranty status is correctly identified as covered
- Accuracy of fault code classification against manuals • target: 0.98 • range: 0.9-1
- Benchmark: BMSFaultBench
- Diagnosis Accuracy target: 0.98
- 1 public reference case
- 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
- Llama Index · Required
- Ray Tune · Optional
- Hume AI · Optional
Evaluation contract
- warranty_coverage_check · Weight 1
- diagnosis_accuracy · Weight 1
Metrics
- diagnosis_accuracy · Target 0.98
Benchmarks
- BMSFaultBench
- energy
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON BMS telemetry log
Output: JSON warranty assessment
Execution harness: Python
Evaluation tools
- LlamaIndex · RAG Framework
Recipe freeze marker
No recipe freeze time is recorded.