Trust and Safety - Senior Data Scientist

Scale AI • San Francisco, CA

11h

About The Position

Scale is at the frontier of Generative AI and human-AI collaboration. The Gen AI Ops Trust and Safety team protects contributor integrity across a marketplace of hundreds of thousands of contributors training foundation models. We're looking for a modeling-focused Data Science Lead who ships fast, thinks in systems, and uses AI coding tools as a core part of their workflow.
This is a high-autonomy IC role. You will own fraud and abuse detection models end-to-end — from label definition through feature engineering, training, evaluation, and production deployment. You'll work in a small team that operates at 10x velocity by pairing deep analytical judgment with AI-augmented development (Cursor, Claude Code). If you've felt limited by teams that move slowly or separate "analysis" from "building," this role eliminates that gap entirely.

Requirements

• 5–8 years in Data Science or Machine Learning, with at least one production fraud/abuse/integrity model shipped end to end (labels → features → deployment → iteration).
• Strong proficiency with AI coding assistants. You should already be using these tools daily and understand how to leverage them for rapid SQL/Python development, data exploration, and code generation — not as a novelty but as a core workflow.
• Deep feature engineering instinct — you see a messy behavioral log and immediately think about what signals it contains, how to extract them, and whether they'll hold up adversarially.
• Expert SQL and Python. Comfortable writing 200+ line CTEs, building data pipelines, and working with large-scale event data.
• Experience with unsupervised methods (clustering, anomaly detection) alongside supervised classification. Bonus if you've worked with semi-supervised approaches for noisy labels.
• Comfort operating across the full stack: Snowflake DDLs, Python modeling, rule engines, alerting systems, Google Docs/Sheets automation, workflow orchestration.
• Clear, direct communication. You can explain a model's precision/recall tradeoff to an ops lead and debug a Snowflake query in the same afternoon.

Nice To Haves

• Marketplace or gig-economy platform experience.
• Experience building linkage/graph-based detection.
• Familiarity with identity verification vendors and their signal taxonomies.
• Experience building automated alerting and monitoring pipelines.

Responsibilities

• Own the model lifecycle — from label definition through feature engineering, training, evaluation, and production deployment. Ground truth is ambiguous and non-binary; you'll need to make it work anyway.
• Build and extend the feature store — design and productize signals across identity, behavioral, and third-party data sources that power detection models reliably at scale.
• Design and ship detection systems — not just notebooks. Rules, clustering, anomaly detection — all the way to production decisions.
• Move fast with AI tools — use AI-assisted IDE and copilots daily to prototype pipelines, run investigations, and automate workflows. This is how the team operates.
• Partner with Ops and Engineering to close the loop between model output and real-world actions, and continuously improve the contributor experience.

Benefits

• Comprehensive health, dental and vision coverage, retirement benefits, a learning and development stipend, and generous PTO.
• this role may be eligible for additional benefits such as a commuter stipend.

What This Job Offers

• Job Type: Full-time
• Career Level: Mid Level
• Education Level: No Education Listed
• Number of Employees: 1,001-5,000 employees