Senior QA Architect for ERP

We're looking for a senior QA Architect to build a testing strategy from scratch for a legacy ERP going through active modernization. IMPORTANT BEFORE APPLYING • FULL AUTONOMY: you'll be the top authority on QA and testing. The team will work alongside you, but no one else will set the strategy, priorities, or decisions. • LEGACY EXPERIENCE: you must have introduced automated testing from scratch in a legacy application with no prior coverage. This is the first filter. • HANDS-ON WORK: beyond recommending, you'll write tests, build PoCs, and implement the first layers together with the team. • TESTING BEYOND THE UI: you need experience designing a full strategy built around business risk and workflows, not just UI checks. • STACK: experience with .NET, web applications, and SQL Server. • LOCATION: based in Europe, with hours overlapping Spain (CET/CEST). • LANGUAGE: professional-level Spanish or higher. • LONG-TERM POTENTIAL: genuine possibility of long-term work if it's a good fit. THE CONTEXT We're a European software company with an ERP made up of around forty modules covering payroll, scheduling, time tracking, invoicing, and accounting. It's a database-heavy system with critical processes and legal implications. Two platforms currently run side by side in production: • LEGACY LAYER: a proprietary 4GL platform, written in FDL and rendered through WinForms on .NET Framework 3.5. Business logic and data access use in-house languages. • MODERN LAYER: a .NET 10 and Blazor Server application backed by SQL Server. A transpiler generates its C# and Razor code every night straight from the legacy system. The legacy system is still the source of truth. The strategic goal is for the modern platform to stop being generated and become directly developable. THE CHALLENGE Business logic has almost no automated coverage. Current tests cover infrastructure and core functionality, but not processes like payroll, invoicing, or scheduling. There's no golden master and no parity strategy to confirm both platforms produce the same results from the same inputs. On top of that, a single transpiler bug can ripple across every module. The job is building the safety net needed to modernize the product with confidence. RESPONSIBILITIES • Assess the architecture, existing testing, risks, and testability constraints. • Define the right mix of characterization tests, golden master, parity, integration, database contract, end-to-end, regression, and exploratory testing. • Design a way to run the same inputs against both platforms and compare business outcomes, database changes, permissions, and side effects. • Define how the transpiler and the generated C# and Blazor business logic get tested. • Structure testing around complete business workflows, not isolated modules or screens. • Increase testability incrementally, without requiring a major refactor upfront. • Define automation, test data, environments, CI/CD quality gates, and reporting. • Set up metrics: critical workflow coverage, production defects, detection time, parity gaps, and release cost. • Build a roadmap, implement the first tests, and transfer knowledge to the team. EXPECTED INITIAL OUTCOMES • Risk and testability assessment grounded in the real system. • Prioritized roadmap covering critical workflows, quick wins, and structural work. • Parity PoC between the legacy system and the generated code. • First business workflows automated and wired into CI/CD. • Metrics and reporting to track coverage, regressions, and production defects. MUST-HAVE • Strong command of testing strategy for complex, database-heavy systems. • Experience with ERPs or other critical applications carrying legal implications. • Ability to work with test data, integrations, database contracts, and end-to-end workflows. • Pragmatic judgment to prioritize by risk and gain coverage despite technical debt. NICE TO HAVE • Experience with 4GL platforms, in-house DSLs, transpilers, or generated codebases. • Experience working natively with AI, OpenSpec, or compound engineering. • Experience building reusable skills for AI agents. TO SUBMIT YOUR PROPOSAL • Describe a case where you introduced automated testing from scratch in a legacy application: starting point, the work you personally did, and the results. • Explain how you'd compare two implementations of the same ERP when one is auto-generated from the other. • Summarize what you'd cover in the first few weeks and what you'd leave out initially. • Share your availability, time zone, and usual rate.