Vacancy details
This is a private vacancy. Company details stay confidential until InVault confirms a serious candidate fit.
- Status
- Active
- Date posted
- 2026-11-15
- Open until
- 2027-02-13
- Employment type
- Full-time
- Source-listed languages
- Not specified
- Language requirement
- No mandatory language inferred
- Work arrangement
- Remote
- Physical workplaces
- Not specified
- Applicant eligibility
- United States; Canada
Overview
Analyze third-party and consortium fraud intelligence datasets to evaluate their real contribution to financial-crime detection. Build structured signal-quality reviews, historical tests and comparison studies for identity and payment risk models, then communicate evidence-based recommendations to fraud, product and data engineering teams.
Requirements
- Three to five years of experience in data analysis, data science or a related analytical role, preferably in fraud, risk, fintech or data-intensive B2B software.
- Strong SQL skills are required, together with Python or R for statistical data manipulation, analysis and visualization.
- Understanding of precision/recall, AUC/ROC, lift, population distributions, A/B-testing fundamentals and false-positive tradeoffs.
- Experience evaluating third-party or external datasets for completeness, freshness, match rates, coverage and signal value.
- Ability to design lift studies, backtests and champion-versus-challenger evaluations of detection signals.
- Strong written and spoken communication translating analytical findings for nontechnical stakeholders.
- Comfort defining evaluation methodology independently in ambiguous environments.
- Specific device identity, consortium or behavioural fraud datasets, vendor-evaluation experience, machine learning and fintech vertical familiarity are advantageous, not mandatory.
Responsibilities
- Design and run evaluation frameworks for external fraud and identity data assets.
- Measure incremental signal value through lift analyses, backtests and comparison tests.
- Profile data quality, coverage, match rates and fraud-signal freshness.
- Recommend adoption, pilot testing, deprioritization or rejection of new data signals based on quantified results.
- Coordinate ingestion and production-representative test data with data engineers.
- Maintain reproducible reporting and historical vendor-signal performance documentation.
- Support fraud investigations and risk-model performance analysis when needed.
Conditions
- Full-time, fully remote role based in the United States or Canada.