The draft to review
“Built an enterprise AI analytics platform that increased revenue by 20% and eliminated data errors. Used Python, SQL, cloud and machine learning.” The only available artifact is the six-order fictional Harbor exercise. Identify which statements are demonstrated, unsupported or misleading.
Your evidence ledger
Create a table locally with claim, artifact, reproducible check and limitation. Require a real artifact for every skill claim: a query for SQL, a runnable script for Python, infrastructure evidence for cloud and a defined evaluation for ML. Listing a tool does not prove it was used. Mark absent evidence as missing, not implied.
Worked review
The available work supports a small SQLite reconciliation and an explanation of join grain. It does not demonstrate enterprise deployment, an AI platform, a causal revenue increase, elimination of all errors, Python, cloud or machine learning. The 20% figure belongs to a different constructed experiment and must not be imported as a project result.
A stronger rewrite
“Created a fictional SQLite revenue-reconciliation case with an explicit cutoff and denominator. Repaired duplicate order value caused by multiple refund rows; reconciled net value to EUR 375 and checked equal-price and extra-refund variants. Documented assumptions, limitations and a short stakeholder explanation.” Add Python/cloud claims only when you have actually built and tested those artifacts.
Reviewer rubric
Score question clarity, source/provenance, reproducibility, controls and honest scope 0–2 each. Each score of two needs a specific artifact or visible example. Treat missing evidence as a next task. A high self-review score is not a certificate, hiring recommendation or guarantee of an interview.
Your revision task
Write a 150-word project summary and a 30-word resume bullet. Ask a peer to underline unsupported claims. Attach a short evidence index naming each file and its purpose, without publishing private keys, personal records or employer-confidential data. Keep this exercise local unless you deliberately choose to publish your own portfolio.
Try the activity before reading its reference. Keep your sources, commands and assumptions with your work. Companion activities are unsaved self-practice and do not award a certificate or change lesson progress.