Reddit Senior Data Scientist at Reddit leading ads fraud detection and scaled enforcement within the Safety organization. Develop detection strategies, analyze large datasets, and build scalable fraud detection pipelines.
Responsibilities
Lead the measurement and detection strategy for ads fraud by defining fraud taxonomies, labels, sampling plans, metrics, and evaluation frameworks that make performance measurable and defensible.
Analyze large, complex datasets and networks of behavior to uncover emerging fraud patterns, size their impact, identify root causes, and translate findings into detection and enforcement requirements.
Design and develop scalable ads fraud detection and enforcement pipelines in partnership with Engineering and Machine Learning, including feature generation, rules and models, near-real-time scoring, actioning, review feedback loops, and observability.
Own the full detection lifecycle: backtesting, threshold calibration, offline and online evaluation, launch validation, experimentation, monitoring, drift detection, incident response, rollback, and retirement.
Build and maintain statistical, machine learning, and GenAI-enabled models or prototypes that improve fraud detection, risk identification, investigator efficiency, and enforcement quality.
Balance fraud loss, platform and advertiser risk, customer experience, false-positive costs, operational capacity, and business goals when recommending detection thresholds and enforcement strategies.
Partner across Ads and Safety to shape strategy and roadmaps, strengthen data foundations, close policy and enforcement gaps, and ensure solutions meet governance and compliance standards.
Translate complex analyses into clear narratives and actionable recommendations for technical and non-technical stakeholders, including senior leaders, and mentor other data scientists and analysts.
Qualification
Relevant experience in Data SciencePhFluency in statistical analysis
Required
Relevant experience in Data Science, Applied Science, or a related quantitative role, preferably in ads fraud, financial fraud, or account risk, Trust & Safety, platform integrity, or enforcement engineering.
Ph.D. or M.S. degree in Statistics, Economics, Computer Science, Applied Mathematics, or another quantitative field; with an M.S., 4+ years of industry data science experience, or with a Ph.D., 2+ years of industry data science experience.
Demonstrated experience building or materially shaping production detection and automated enforcement pipelines, including batch or streaming data, feature engineering, rules or models, decisioning, monitoring, and feedback loops.
Strong command of fraud or abuse detection methods and evaluation, including label design, precision and recall tradeoffs, calibration, threshold selection, false-positive analysis, drift detection, and adversarial adaptation.
Experience partnering closely with Product and Engineering teams to translate analyses and prototypes into reliable production systems; experience working across Ads, Safety, fraud, risk, or platform-integrity organizations is preferred.
Experience applying AI and large language models (LLMs) to practical data science workflows, such as threat discovery, content classification, signal development, investigation automation, or detection and enforcement systems.
Deep understanding of complex behavioral networks or large-scale activity patterns; experience with methods such as graph or network analysis, clustering, anomaly detection, or natural language processing is valuable.
Fluency in statistical analysis, Python or a similar programming language, and SQL, with the ability to work independently across complex data systems and unfamiliar codebases.
Ability to tackle ambiguously defined problems, deconstruct them into precise and tractable components, and move from investigation to scalable, reusable solutions.
Strong technical leadership and communication skills, with a track record of influencing cross-functional roadmaps, aligning stakeholders, and explaining complex topics to technical and non-technical audiences.
