R

Riskified

eCommerce fraud prevention and revenue protection

Staff Data Scientist: Foundation Models

Applied Machine LearningPosted on Nov 16
Full-timeTel Aviv-Yafo, Israel

About the Job

Requirements

  • A Master of Science or Ph.D. in Computer Science, Statistics, Mathematics, or a related quantitative field.
  • At least 6 years of hands-on experience designing, training, and deploying machine learning models in large-scale production environments.
  • Hands-on experience building, fine-tuning, or adapting foundation models or similar large-scale models, such as pre-trained transformers, self-supervised sequence models, tabular foundation models, or large language models, and sound judgment on when to build versus reuse.
  • A track record of turning open-ended research questions into well-designed experiments and evidence-based decisions.
  • Strong practical experience with large-scale distributed data processing, such as Spark or Databricks, and training models on large datasets.
  • Proven technical leadership beyond one's own team, including setting direction, aligning stakeholders, and mentoring senior scientists.
  • Ability to explain complex research and trade-offs clearly to technical and non-technical audiences, including senior leadership.
  • Ability to learn unfamiliar technology, solutions, data, and domains quickly.

Responsibilities

  • Lead the research and roadmap for applying foundation models to core fraud-prevention problems, from proof of concept through real-time production.
  • Challenge assumptions, set success criteria tied to business metrics, and use early experimental evidence to guide decisions before scaling.
  • Choose whether to train proprietary models or adapt existing models, then develop effective sequence, transformer, graph, or deep tabular models for large-scale transaction and behavioral data.
  • Learn the fraud domain, data, and production models, and identify where learned representations can add value.
  • Partner with Data Engineering, ML Platform, Core Software, and Product on the data, training, and serving infrastructure required by the models.
  • Bring current research into practical solutions, set standards for research rigor, and mentor data scientists across the department.

Desired Qualifications

  • Experience pre-training models on event, transaction, or behavioral sequences involving non-text data.
  • Experience with tabular foundation models, graph neural networks, or embedding-based models at scale.
  • Experience in fraud detection, risk management, payments, or fintech.
  • Relevant publications, open-source work, or talks.

About the Company

Riskified

eCommerce fraud prevention and revenue protection

Riskified helps online merchants protect revenue and prevent fraud in ecommerce by using machine-learning models to approve or block transactions in real time before they reach the bank. Its products include pre-authorization decisions and Dynamic Checkout that adjusts the checkout experience based on each shopper’s risk level, reducing friction for legitimate customers while flagging suspicious activity. The company differentiates itself with real-time, pre-authorization ML decisions that tie fraud protection directly to revenue growth and a checkout experience that adapts to risk, scalable for merchants expanding into new markets. Its goal is to help merchants grow revenue and improve customer satisfaction by preventing fraud, lowering false declines, and streamlining the digital checkout process.

Company size
501-1,000
Company stage
IPO
Headquarters
New York City, New York
Founded
2013

Recently Posted Jobs

Explore more opportunities at Riskified

See more jobs