Fraud-Focused Foundation ML Scientist (Remote)

Sardine; California (MO); On-site; USD 150,000 - 230,000; Full time

Benefits offered by this job

Remote-first culture

Health insurance

401k matching

Job summary

Sardine is seeking an experienced ML applied scientist to advance foundation-model-based fraud detection across real-time platforms. You will work with rich, non-text sequential data to design, train, and deploy state-of-the-art models that operate at scale in production environments.

This role emphasizes practical experimentation, robust evaluation, and close collaboration with customers. You will partner with engineering, data science, and product teams to push the boundaries of fraud

Qualifications

• 4+ years in applied machine learning, quantitative modeling, or ML engineering including at least one foundation model pre-trained or adapted

Responsibilities

• Identify and scope opportunities, design experiments, and execute foundation model research on fraud detection
• Own evaluation bar for foundation model performance including offline benchmarks and head-to-head comparisons
• Take models through data prep, pretraining, fine-tuning, distillation, and deployment with low-latency inference
• Collaborate with Engineering on training infra, GPU efficiency, feature stores, and production serving
• Work with client-facing teams to translate model capabilities into actionable decisions for risk teams
• Coordinate with Legal and Compliance for explainability and governance in regulated environments

Skills

Foundation models

Python

SQL

Model serving

Experiment design

Job description

Sardine is seeking an experienced ML applied scientist to advance foundation-model-based fraud detection across real-time platforms. You will work with rich, non-text sequential data to design, train, and deploy state-of-the-art models that operate at scale in production environments.

This role emphasizes practical experimentation, robust evaluation, and close collaboration with customers. You will partner with engineering, data science, and product teams to push the boundaries of fraud