Acceptance Harness

Portfolio positioning: Senior Data Platform Engineer | Governed Ingestion, Azure, Kafka, Microsoft Graph

This infrastructure-independent harness turns message-processing requirements into deterministic fixtures, typed outcomes, contract checks, and CI gates so release decisions do not depend on production systems.

Role signal: A hiring manager should infer disciplined ownership of acceptance criteria, deterministic test design, contract evolution, failure routing, and release confidence.

Evidence boundary: Deterministic validation and contract testing are implemented here. Azure, Microsoft Graph, and executable replay/idempotency behavior are ?? VERIFY; they are not evidenced by this repository.

What This Proves

Architecture or Control Flow

1. FixtureLoader reads synthetic messages from fixtures/messages.json.
2. FixtureValidator and MessageProcessor validate sender, payload, and correlation-ID inputs.
3. Injected clock and ID providers remove time- and randomness-based test drift.
4. Processing returns a typed accepted or rejected result with a stable failure code.
5. A domain-event publisher exposes decision evidence without requiring an external broker.
6. Maven tests and privacy/claim scanning fail the CI pipeline when behavior or public-safety rules regress.

This validates application and contract behavior in isolation. Kafka offsets, database transactions, production latency, and deployed replay behavior remain outside this harness's proof boundary.

Verification

PowerShell, from the repository root:

Push-Location acceptance-harness
.\mvnw.cmd -B test
Pop-Location

python -m pytest tests/test_pii_scan.py -q
python tools/redaction/pii_scan.py README.md acceptance-harness/README.MD acceptance-harness/ACCEPTANCE_CRITERIA.md acceptance-harness/RUNBOOK.md acceptance-harness/src docs

Bash, from the repository root:

cd acceptance-harness
chmod +x mvnw
./mvnw -B test
cd ..

python -m pytest tests/test_pii_scan.py -q
python tools/redaction/pii_scan.py README.md acceptance-harness/README.MD acceptance-harness/ACCEPTANCE_CRITERIA.md acceptance-harness/RUNBOOK.md acceptance-harness/src docs

Evidence

Platform Engineering Portfolio

These repositories form one evidence-based platform engineering portfolio:

• airgap-messaging-ingest - event-driven ingestion, replay safety, and observability
• acceptance-harness - deterministic validation and release confidence
• ai-platform-operating-model - governed AI platform controls and review-first operations

Repository Layout

acceptance-harness/
  src/main/java/com/example/processor/domain/
  src/main/java/com/example/processor/application/
  src/main/java/com/example/processor/fixtures/
  src/test/java/com/example/processor/
  fixtures/messages.json
  contracts/
  docs/adr/
  ACCEPTANCE_CRITERIA.md
  RUNBOOK.md
docs/
  PROOF_RECEIPTS.md
tools/redaction/
tests/

Reviewer Path

1. Run the PowerShell verification block.
2. Inspect acceptance-harness/src/main/java/com/example/processor/application/MessageProcessor.java as the decision boundary.
3. Inspect the two JSON schemas under acceptance-harness/contracts/.
4. Review the ADRs under acceptance-harness/docs/adr/ for the documented tradeoffs.
5. Inspect JUnit output under acceptance-harness/target/surefire-reports/ after the local run.

Safety and Non-Claims

This is a synthetic/demo implementation. It does not claim formal compliance, production deployment, client-private outcomes, infrastructure-level Kafka/database behavior, or measured production latency.

This repository does not claim IL4/IL5 certification, ATO approval, RMF approval, STIG compliance, or DoD production deployment. CI publishes JUnit reports only as CI artifacts; raw private exports are not published.