The setting

Voltbarn installs home solar and battery systems, and every unit streams telemetry back to the platform. You are a data engineer on the platform team, responsible for the pipelines that turn raw device readings into the numbers operations and billing depend on.

Every mission on this path happens at the same company, so context carries over the way it does in a real job: the data you cleaned in mission two is the data the finance lead questions in mission four.

The missions

1. pipeline designstarter

Design the ingestion pipeline for Voltbarn's new battery telemetry

Voltbarn is shipping a new battery line, and the platform team needs an ingestion pipeline before the first units go live. The old pipeline was built for a smaller fleet and one firmware version, and it will not hold.

You deliver: A pipeline design document covering ingestion, buffering, storage, error handling, and how it scales to the stated volumes.

Scored on: Sizes the stream, Normalizes at ingest, Reliability and idempotency, Meets the targets.

Working from: ingest_requirements.md, telemetry_sample.csv.

2. schema modelingstarter

Model the warehouse schema for Voltbarn installs and readings

The raw Voltbarn tables were copied straight from the operational systems, and every dashboard query against them is slow and wrong in a different way. The team wants a clean modeled warehouse to build on.

You deliver: A dimensional schema design with fact and dimension tables, grains, keys, and slowly changing dimension decisions.

Scored on: Correct fact grain, Dimensions identified, SCD decision, Keys and foreign keys.

Working from: raw_tables.md.

3. ETL checkscore

Write the data quality checks for Voltbarn's nightly readings load

Voltbarn's nightly readings load has been passing bad data straight to billing, and last week a customer was billed off a sensor that reported an impossible reading. The team wants real checks before the next run.

You deliver: A data quality check suite plus a report of exactly which sample rows fail which check.

Scored on: Covers the contract, Catches the real errors, Avoids false positives, Severity and action.

Working from: readings_load_sample.csv, load_contract.md.

4. orchestrationcore

Fix the orchestration behind Voltbarn's nightly warehouse build

On the night of March 12 the source API was slow, and Voltbarn's executive dashboard showed half a day of readings while billing exported off incomplete data. The jobs technically all ran, which is the problem.

You deliver: An orchestration design with a corrected dependency graph, trigger and retry policy, and a fix for the March 12 failure.

Scored on: Fixes the dependencies, Retry and backfill policy, Correct root cause, Freshness gate.

Working from: jobs.csv, incident_log.md.

5. lineage and monitoringstretch

Trace Voltbarn's missing kilowatt-hours and build the monitoring plan

Finance says Voltbarn's dashboard energy total for March runs about 12 percent below what the meters report, and it has been wrong long enough that nobody trusts the number. You need to find where the energy goes and make sure it never silently vanishes again.

You deliver: A lineage trace from the dashboard energy total back to the source, plus a monitoring plan with checks, thresholds, and owners.

Scored on: Traces the lineage, Finds the row drop, Correct fix, Monitoring plan.

Working from: pipeline_map.md, dim_gap_notes.md.

How the scoring works

Each deliverable is graded against the rubric written for that mission. Separately, every mission on every path is graded on how you used AI, against the same four criteria:

  • Understood the task. The learner framed the goal for the assistant clearly instead of pasting the brief and hoping.
  • Grounded in the material. The learner directed the assistant into the provided files and based the work on them, not on invented facts.
  • Verified the output. The learner checked claims, numbers, or coverage against the source material before submitting.
  • Iterated with judgment. The learner refined weak parts of the draft with specific follow-ups rather than accepting the first answer.

Both scores, with the work behind them, go on your proof profile. That is what makes a claim like "I can use AI for data engineering" something an employer can check.