Senior Data Scientist - Alt Defense

WHY DATA SCIENCE & ANALYTICS?

The Data Science & Analytics organization's mission is to increase our speed, frequency and acumen of making decisions at scale by instilling a data-influenced approach to building products. We cover a wide area of the data spectrum including analytical data engineering, product analytics, experimentation, causal inference, statistical modeling and machine learning. Aligned and partnering with product verticals, we use this extensive tool belt to discover new opportunities and unmet use cases, influence and shape the product roadmap and prioritization, build data products and measure impact on our community of players and developers.

WHY ALT DETECTION?

Our Alt Detection system at Roblox is a frontier challenge in large-scale identity modeling, managing a massive multi-partite graph that currently includes approximately tens of billions of nodes and hundreds of billions of edges. In this role, you will apply your expertise in data science, statistics, and causal inference to strategically define, measure, and improve our alt detection systems. This is integrated deeply within various applications across the company. Within our Safety group, identity modeling is critical to mitigate alternate accounts used by adversarial actors to bypass our safety enforcements. Beyond safety, your work will be critical to systemic growth, used for core growth reporting and understanding true market penetration, in addition to establishing fair creator incentives. And more opportunities to expand this system across safety and non-safety use cases. This is an opportunity to build state-of-the-art detection systems that fuel our safety efforts while directly catalyzing our company’s most important growth levers.

You Will:

• Develop ground truth evaluation strategies by integrating diverse signals, ranging from "gold standard" manual labels to automated weak supervision frameworks.


• Calibrate score thresholds across applications, balancing high-precision enforcement needs with high-recall required for growth reporting


• Partner with MLE to identify and validate new behavioral, network, and biometric signals to expand detection coverage.


• Collaborate with Engineering to design modular systems capable of processing billions of candidate pairs and supporting high-concurrency graph traversals.


• Lead the cross-functional integration of identity models into Discovery, Ads, and Creator Success to eliminate fraud and improve the fidelity of platform metrics.


• Conduct strategic research to decode the incentives behind alt creation and develop sophisticated classification taxonomies to guide mitigation strategies.


• Communicate strategic insights and present recommendations to leadership and all cross-functional partners, translating complex statistical findings on prevalence, ground truth quality, and effectiveness into actionable strategies for Product, Engineering, Policy, Compliance, and Legal.


• Partner with ML and Data Engineering teams to ensure model development, reporting, and detection systems are built on statistically sound ground truth and measurement frameworks.

You Have:

• 8+ years of industry experience in data science, economics, analytics, or machine learning engineering


• 6+ years of experience using scripting languages (Python, R), and big data query/processing languages and tools such as SQL, Hive, Spark, and Airflow


• Knowledge of ML and Deep Learning either via formal training or industry experience


• Ability to apply creative first-principles reasoning to solve ambiguous problems


• Experience developing large-scale safety or moderation systems as well as experience with content platforms, specifically user-generated content


• Advanced Degree and/or PhD in Statistics, Computer Science, Physics, Applied Math, Economics, or other related quantitative fields