Data Validaton Analyst

Back to all rolesData Validation AnalystApplyJob detailsEngineering/ Product/ TechTiranaFull-timeAbout Cardo AICardo AI builds AI-powered infrastructure for the private debt and structured finance markets, helping asset managers, lenders, and investors manage data, monitor portfolios, and streamline deal operations. Our Integration team sits at the core of this mission, turning complex, unstructured client data into clean, reliable, platform-ready information.About the Role Cardo AI is looking for a Data Validation Analyst to join our Integration team, the group responsible for onboarding structured finance deals spanning consumer lending, CRE, equipment & infrastructure, and other asset classes onto our platform. As part of our deal onboarding process, we use an in-house AI agentic workflow to pull unstructured data directly from complex financial documents (credit agreements, servicing reports, deal cover pages, and loan tapes).This is not a traditional QA role. You will own the quality layer between what the AI extracts and what our clients rely on, and your mandate is to progressively automate yourself out of the manual work. You'll start by validating AI output hands-on, but the goal is to turn what you learn into AI agents that run the checks for you: automated reconciliation agents, anomaly detectors, deal-type-specific validation suites. Every manual check you do is a candidate for an agent you build.You'll work at the intersection of structured finance and applied AI, with a direct line to the technology team: your findings shape how our models evolve, and your agents become part of the product.What You'll Do Validate and understand. Check AI-extracted data against source documents (contracts, servicing reports, loan tapes, cover pages) for your assigned deals. This is how you learn where the pipeline is strong, where it breaks, and what "accurate" means for each deal type.Build agents that do the checking. Turn recurring validation patterns into AI agents and automated workflows, reconciliation agents that compare extractions to source documents, anomaly detectors that flag drift, validation suites tailored to each asset class. You define the logic; you don't need to be a software engineer to build them.Improve the pipeline itself. Propose and prototype new features and functionalities for the extraction and validation process, better confidence scoring, smarter sampling strategies, new check types for new deal structures. Package model performance insights into structured feedback that directly shapes how the technology team evolves the models.Define quality for new territory. Partner with Integration engineers on new deal onboarding, establishing what "accurate" looks like for new fields, asset classes, and data structures, then encode that definition into automated checks.Keep leadership informed. Report on data quality status and automation coverage as deals move through onboarding.What We're Looking For •      1 year of experience in data analysis, QA/data validation, financial data operations, or a related analytical role.•      Strong attention to detail and comfort digging into granular, sometimes messy, data to find discrepancies.•      Working knowledge of SQL (or willingness to learn quickly); Trino/dbt experience is a plus but not required.•      Comfort with spreadsheets (Excel) for reconciliation and reporting; ability to build clear, repeatable QA workflows.•      Interest in structured finance and willingness to build domain knowledge across asset classes like consumer lending, CRE, and specialty finance, no prior structured finance background required, though it's a plus.•      Clear written communication skills you'll be translating what you find into feedback that a technology team can act on.•  Genuine interest in AI/LLM tools you've experimented with them, you're curious about how they fail, and you're excited by the idea of building with them, not just checking them. Prompt engineering or...