# Onboarding validation work sample
Prepared for Gunel Abasova. Independent demonstration with fictional requirements and synthetic data; not an employer deliverable.

## Reproduce
Java 17 or later: `java OnboardingChecks.java`
Python 3: `python3 run_sql.py`
No external libraries are required for these executed checks.

Expected Java result: seeded baseline 6 pass / 2 fail; corrected model 8 pass / 0 fail.
Expected SQL result: one seeded mismatch, zero mismatches in the corrected fixture.

## Evidence status
- Java: eight checks executed against an in-memory reference model. This is NOT an executed TestNG/Cucumber suite or browser test.
- SQL: executed against temporary SQLite fixtures, separate from the Java model.
- Gherkin: specification sample, not wired to step definitions in this package.
- Postman: importable collection design sample, not run. It requires an authorized sandbox API implementing the fictional contract, a fresh C01 fixture, and manual cleanup. No production endpoint is configured.
- Employer experience: user-supplied career description, separate from this sample.

## Requirements
R01: accept a unique, nonblank ID, a basic email format, and US/CA/GB country codes. Invalid input must not create a record (T01–T04, T08).
R02: reject duplicate IDs without modifying the original record (T05).
R03: flagged requests enter manual_review; others enter pending (T01, T06).
R04: preserve submitted values (T07 and SQL fixture check).
These are demonstration rules, not legal, identity-verification, or compliance standards.

## Seeded defects
D01: duplicate ID overwrites an existing record. High impact; block the affected workflow until corrected. T05 reproduces rejection failure.
D02: review flag is ignored and record enters pending. High impact under the fictional rule; T06 reproduces the incorrect state.
The defects were deliberately introduced, not found in an employer product.

## Release decision
Ready for demonstration review only. Production would require authenticated authorization, persistence, concurrent-write protections, full API integration, browser tests, privacy review, and stakeholder UAT. The basic email regex is deliberately limited and is not comprehensive address validation.

## How to discuss this work
Read and run the checks before presenting. Explain why a success status alone is not enough, why submitted data must be compared with retrieved data, and why eight green checks do not prove production readiness.

## Creation note
This sample and website were developed with AI assistance. Review, personalize, and take responsibility for the decisions you present. Do not describe the generated artifacts as confidential employer work or claim unsupported production outcomes.