Developer Needed — Automated Regression Test Harness for Fraud Detection Engine

Project Overview
We have an in-house fraud/risk detection system that scores signals like proxy usage, VPN usage, device fingerprint risk, bot probability, and location spoofing (similar to tools like Fingerprint/Seon). We need an automated test harness that validates our detection engine against known-labeled test cases, so we catch regressions every time we update our rules or models.

What This Is NOT
This is not live evasion testing against production. All test data is pre-labeled, synthetic, and sourced from public reference data (known proxy/VPN IP ranges, published emulator fingerprint signatures, etc.). No real user data, no production access required.

Scope of Work

Build a fixture library of labeled test cases across categories: proxy, vpn, device_risk, location_spoofing, bot, and clean (control cases expected to score low)
Build a test runner that feeds each fixture to our detection API and compares actual output against expected output
Track both false negatives (missed risk) and false positives (over-flagged clean cases) separately
Integrate into our CI pipeline so it runs automatically on every PR/deploy
Output a clear pass/fail summary report per category

Requirements

Experience building automated test suites / CI pipelines
Comfortable working with JSON fixtures and REST APIs
Familiarity with fraud/risk detection concepts is a plus, not required
Clean, documented, maintainable code (this becomes a long-term internal QA tool)

Deliverables

Fixture library (JSON, ~15-20 cases per category to start, extensible)
Test runner + CI integration
Summary report format (pass/fail, false positive vs false negative breakdown)
Short doc on how to add new fixtures going forward