Data Scientist ll - RiskOS

Socure’s RiskOS team is building workforce verification capabilities to reduce hiring fraud and strengthen identity trust in real-world workflows. In this role, you will lead the end-to-end data science lifecycle for workforce identity and hiring risk use cases, working across multi-source data to uncover fraud signals and operationalize them as rules, conditions, machine learning, and GenAI components.

This is an onsite position in New York, NY, with a salary range of USD 140,000 - 170,000 per yearly. You will start with 3+ years of relevant experience and a quantitative academic foundation (or equivalent practical experience) to support modeling, evaluation, deployment, and monitoring within RiskOS.

What you will do

• Own the full data science lifecycle for Workforce Verification on RiskOS, from data exploration and hypothesis generation through model development, evaluation, deployment, and monitoring.
• Explore and analyze workforce-related data sources, including applications, resumes, device and behavioral telemetry, background checks, and ATS/HRIS integrations, to identify patterns of workforce fraud such as fake resumes, identity rental, deepfake interviews, and injection attacks.
• Design, implement, and iterate on rules, conditions, and heuristic logic inside RiskOS workflows to detect high-risk workforce events, including repeated identities across multiple resumes, suspicious device patterns, and anomalous hiring flows.
• Develop and evaluate machine learning models for workforce risk and identity assessment, such as fraud risk scoring, clustering related identities, and anomaly detection across hiring funnels, leveraging Socure identity and device signals when relevant.
• Collaborate with RiskOS and Workforce product teams on GenAI-powered features (including a Resume Verification Agent and explanation agents) to define inputs/outputs, build evaluation datasets, and establish quantitative and qualitative evaluation frameworks for LLM-based components.
• Work closely with engineering to productionize models, rulesets, and GenAI components in RiskOS by defining interfaces, supporting integration and testing, and contributing to monitoring, alerting, and feedback loops.
• Translate model and rules performance into clear customer narratives for product, Workforce GTM, and solution consulting, including impact such as blocking fake applicants, reducing deepfake interviews, or preventing identity rental in hiring.
• Use customer feedback and outcome data to continuously improve Workforce Verification logic and models, including support for experimentation and offline test harness design to safely evaluate new workflows and templates.
• Operate with a product mindset by documenting assumptions and decisions, communicating trade-offs clearly, and proactively surfacing risks, limitations, and improvement opportunities.

What you bring

• Bachelor’s or Master’s degree in a quantitative field such as Computer Science, Statistics, Mathematics, Engineering, or equivalent practical experience.
• 3–6 years of hands-on experience in data science, machine learning, or applied analytics, with meaningful experience preferred in fraud, risk, trust & safety, or workforce/hiring analytics.
• Experience owning end-to-end analytics and/or model development, including problem framing, data wrangling, feature engineering, model training, evaluation, and deployment support.
• Strong proficiency in Python and SQL, including common data science and ML libraries such as pandas, scikit-learn, XGBoost, PySpark, or similar.
• Ability to work with large, messy, heterogeneous datasets (for example, JSON workflows, logs, event streams, and third-party enrichments), building reusable abstractions or utilities for yourself and others.
• Exposure to Natural Language Processing and/or unstructured text analytics, such as resume or document parsing, entity extraction, similarity search, or basic embedding-based methods, ideally in real product contexts.
• Some hands-on experience with Generative AI / LLM-based products, including commercial LLM APIs, prompt design, RAG-style retrieval, or evaluation of LLM outputs, plus interest in deepening the skill set.
• Strong analytical and problem-solving skills, including comfort reasoning about ambiguous signals and adversarial behavior in fraud or workforce contexts.
• Willingness to take on light data engineering and production-oriented tasks when needed in partnership with engineering, such as building ETL transforms, contributing to Airflow/Spark jobs, or instrumenting basic monitoring.
• Clear, concise communication skills, including the ability to explain complex analyses, models, and GenAI behavior to non-technical stakeholders (product, GTM, customers).
• A bias toward ownership, learning, and collaboration, comfortable receiving guidance from senior data scientists while expanding scope and autonomy.

Tools you will use

Python, SQL, pandas, scikit-learn, XGBoost, PySpark, Natural Language Processing, Generative AI, LLM-based products, commercial LLM APIs, RAG-style retrieval, Airflow, Spark

Nice to have

• Direct experience with workforce, HR tech, ATS/HRIS data, or hiring funnel analytics.
• Prior work on identity verification, device intelligence, or orchestration/rules engines (for example, RiskOS or similar systems).
• Familiarity with evaluation and monitoring of GenAI systems, including offline benchmarks, human-in-the-loop review, and safety or hallucination checks.