Staff Data Scientist

About the role

Socure is actively searching for a technical leader to join the Digital Intelligence team with the specific mandate to transform complex telemetry into production-grade identity and fraud signals. The successful candidate will own the end-to-end strategy for detecting risky behaviors and verifying trusted devices while providing technical direction through ambiguous and high-stakes challenges. This role requires a balance of deep analytical rigor and pragmatic execution to ensure that models serve both security efficacy and business stability. You will be responsible for defining the evaluation frameworks that determine whether a signal is robust enough to ship to production. The position involves close collaboration with product, engineering, and operations to ensure that data strategies align with real-world constraints. You will investigate sophisticated adversarial behaviors including spoofing, automation, and evolving VPN tactics. Success in this role is defined by the ability to mentor and elevate the entire data science organization's modeling judgment and standards.

Key facts

What you'll do

• Lead machine learning initiatives that span device, network, browser, mobile, and behavioral intelligence domains.
• Architect and deploy risk signals that carefully balance fraud detection accuracy with false positive control and operational stability.
• Design and implement rigorous evaluation methods, including drift monitoring, leakage checks, and adversarial robustness testing.
• Partner closely with engineering and product teams to influence telemetry collection strategies, data contracts, and production readiness criteria.
• Investigate complex behavioral patterns such as spoofing, automation, VPN usage, and the challenges posed by device fragmentation.
• Mentor data scientists to improve modeling judgment, validation standards, and the clarity of technical documentation.
• Define the metrics and experiments that prove the incremental value of new signals before and after deployment.
• Translate ambiguous business problems into well-defined data science hypotheses with clear success criteria.
• Ensure that models remain reliable over time by monitoring data quality, feature integrity, and prediction drift.
• Act as the technical authority for risk modeling within the Digital Intelligence team and across the broader organization.

Requirements

• Hold a Master's or Ph.D. in Computer Science, Machine Learning, Statistics, Mathematics, or a related quantitative field.
• Bring 12 or more years of hands-on experience in data science, applied machine learning, or statistical modeling.
• Demonstrate significant experience deploying and validating production-grade risk signals or decisioning systems at scale.
• Possess a strong background in fraud detection, cybersecurity, trust and safety, or anomaly detection within digital ecosystems.
• Show expert-level SQL skills combined with proficiency in Python and distributed processing frameworks such as Spark or PySpark.
• Prove a history of handling imperfect labels, telemetry gaps, and shifting adversarial patterns without model degradation.
• Exhibit strong judgment in model evaluation, including the selection of appropriate metrics and validation strategies.
• Communicate effectively with both technical and non-technical stakeholders to align on priorities and constraints.

Nice to have

• Hands-on experience with device fingerprinting, behavioral biometrics, or graph-based risk signals.
• Familiarity with high-cardinality categorical data, embeddings, or representation learning techniques.
• Knowledge of streaming architectures or low-latency decisioning systems that support real-time risk scoring.
• Experience with frameworks such as scikit-learn, XGBoost, TensorFlow, or PyTorch in production environments.
• Background in adversarial, interpretable, or privacy-preserving machine learning methods.
• Prior work in regulated environments where model explainability and auditability are critical.

Skills & tools

• Python, SQL, Spark, PySpark, scikit-learn, XGBoost, TensorFlow, PyTorch.

Practical notes

• Socure is an equal opportunity employer. If you require accommodations during the application or interview process, please contact your recruiting partner.

About the role

Socure is actively searching for a technical leader to join the Digital Intelligence team with the specific mandate to transform complex telemetry into production-grade identity and fraud signals. The successful candidate will own the end-to-end strategy for detecting risky behaviors and verifying trusted devices while providing technical direction through ambiguous and high-stakes challenges. This role requires a balance of deep analytical rigor and pragmatic execution to ensure that models serve both security efficacy and business stability. You will be responsible for defining the evaluation frameworks that determine whether a signal is robust enough to ship to production. The position involves close collaboration with product, engineering, and operations to ensure that data strategies align with real-world constraints. You will investigate sophisticated adversarial behaviors including spoofing, automation, and evolving VPN tactics. Success in this role is defined by the ability to mentor and elevate the entire data science organization's modeling judgment and standards.

What you'll do

• Lead machine learning initiatives that span device, network, browser, mobile, and behavioral intelligence domains.
• Architect and deploy risk signals that carefully balance fraud detection accuracy with false positive control and operational stability.
• Design and implement rigorous evaluation methods, including drift monitoring, leakage checks, and adversarial robustness testing.
• Partner closely with engineering and product teams to influence telemetry collection strategies, data contracts, and production readiness criteria.
• Investigate complex behavioral patterns such as spoofing, automation, VPN usage, and the challenges posed by device fragmentation.
• Mentor data scientists to improve modeling judgment, validation standards, and the clarity of technical documentation.
• Define the metrics and experiments that prove the incremental value of new signals before and after deployment.
• Translate ambiguous business problems into well-defined data science hypotheses with clear success criteria.
• Ensure that models remain reliable over time by monitoring data quality, feature integrity, and prediction drift.
• Act as the technical authority for risk modeling within the Digital Intelligence team and across the broader organization.

Requirements

• Hold a Master's or Ph.D. in Computer Science, Machine Learning, Statistics, Mathematics, or a related quantitative field.
• Bring 12 or more years of hands-on experience in data science, applied machine learning, or statistical modeling.
• Demonstrate significant experience deploying and validating production-grade risk signals or decisioning systems at scale.
• Possess a strong background in fraud detection, cybersecurity, trust and safety, or anomaly detection within digital ecosystems.
• Show expert-level SQL skills combined with proficiency in Python and distributed processing frameworks such as Spark or PySpark.
• Prove a history of handling imperfect labels, telemetry gaps, and shifting adversarial patterns without model degradation.
• Exhibit strong judgment in model evaluation, including the selection of appropriate metrics and validation strategies.
• Communicate effectively with both technical and non-technical stakeholders to align on priorities and constraints.

Nice to have

• Hands-on experience with device fingerprinting, behavioral biometrics, or graph-based risk signals.
• Familiarity with high-cardinality categorical data, embeddings, or representation learning techniques.
• Knowledge of streaming architectures or low-latency decisioning systems that support real-time risk scoring.
• Experience with frameworks such as scikit-learn, XGBoost, TensorFlow, or PyTorch in production environments.
• Background in adversarial, interpretable, or privacy-preserving machine learning methods.
• Prior work in regulated environments where model explainability and auditability are critical.

Practical notes

• Socure is an equal opportunity employer. If you require accommodations during the application or interview process, please contact your recruiting partner.