About the role

We are modernizing Cascade, a large, business-critical auction and asset-management platform, and are in the business testing and data migration phase. The legacy system holds years of seller configuration, workflow, fee and rules data. Much of its business logic lives in stored procedures, lookup tables and application code rather than in documentation.

We need a hands-on Database SME who can quickly understand a complex legacy SQL Server estate, produce accurate data extracts for business teams, find and fix data-quality problems, and support migration to the modernized platform. The timeline is short. We're looking for someone who uses LLM-based tools (Claude, Copilot, ChatGPT and similar) as a normal part of their daily work to read unfamiliar code, write and review complex queries, and move faster than a traditional approach allows, while staying rigorous about accuracy.

What you'll do

- Reverse-engineer legacy business logic. Trace how screens get their data, through ASP.NET pages, C# data-access layers (LINQ to SQL / Entity Framework), stored procedures, functions and lookup tables. Turn that into explicit, reliable SQL.

- Build data extracts for business users, such as Finance and Product. The extracts must match what the application shows, including its defaults, hidden fields, versioning rules and quirks, and must be easy for non-technical readers to use in Excel.

- Write complex, read-only production queries safely. Use window functions, version resolution ("current effective record"), rewriting procedure logic as plain queries where execute rights aren't available, and set-based work across large tables, without blocking or disturbing production.

- Find and quantify data-quality issues. Look for duplicates, orphaned records, stale versions and inconsistent lookup codes. Document them and propose cleanup scripts for review and approval.

- Support data migration. Map legacy schemas to the modernized data model,



write and check migration and reconciliation queries, and confirm row counts and values match between old and new systems.

- Reconcile against the UI. Work with business testers to compare extracts with what the legacy and modernized screens show, and explain any differences.

- Document as you go. Record business rules, field-to-column mappings and known data problems, so the knowledge outlasts the contract.

Must have

- 8+ years of hands-on SQL Server development: T-SQL, stored procedures, functions, window functions, CTEs, APPLY, temp tables, execution plans. Experience with Azure SQL / Managed Instance preferred.

- A proven record of understanding large, poorly documented legacy databases quickly, and pulling the rules out of code and data.

- Enough C# / .NET to trace data access through application code: ASP.NET WebForms or MVC, LINQ to SQL, Entity Framework.

- Hands-on, daily use of LLM coding tools in your current work: reading code with AI assistance, writing queries and scripts, reviewing output critically, and checking results against data. You should be able to walk us through a concrete example.

- A disciplined, verify-everything approach. You check AI-generated SQL against real data and never ship a number you haven't confirmed.

- A safe-by-default attitude toward production: read-only by default, reviewed changes, and no shortcuts on shared environments.

- Experience with data migration and reconciliation projects.

- The ability to explain technical findings clearly to non-technical stakeholders, e.g. why an extract shows a field or doesn't.

Nice to have

- Domain experience in real estate, mortgage servicing, REO, foreclosure or auction platforms.

- Experience building Excel-ready outputs for business teams: layout, formatting, pivot-friendly structures.

- Familiarity with SQL Server database projects (SSDT / .sqlproj), Git, and Azure DevOps.

- Python or PowerShell for automating repetitive data tasks.

📌 DBA (Bengaluru)
🏢 Kadel Labs
📍 Bengaluru