Remote Senior Anti-Fraud ML Engineer

Toloka; United States; Hybrid; USD 120,000 - 180,000; Full time

Benefits offered by this job

Remote or Hybrid work model

Competitive compensation

ESOP

Paid PTO

Home office allowances

Job summary

Toloka AI is building fraud-detection systems at the intersection of analytics and applied ML. You will analyze data to uncover patterns, generate hypotheses, and develop defenses—rules, models, and LLM-based approaches—deployed in production.

You’ll own end-to-end delivery from insights to monitoring, and you’ll build data pipelines (SQL, Python, Airflow), design production metrics, and contribute to anti-fraud policy shaping in a remote-friendly hybrid setup.

Qualifications

• Strong SQL skills and ability to query large data sets (ClickHouse a plus).
• Proficiency in Python and ML techniques; experience deploying models to production.
• Experience building data pipelines with Airflow or similar.

Responsibilities

• Explore data to identify fraud patterns, anomalies, and insights.
• Design metrics and build dashboards to visualize protection effectiveness.
• Generate hypotheses from data and rigorously test them.
• Turn hypotheses into production solutions: rules, ML models, and LLM-based approaches.
• Build and maintain data pipelines (SQL, Python, Airflow) to keep defenses up to date.
• Think like an attacker and design defenses to withstand day-one bypass attempts.
• Own area end to end from insight to deployment and monitoring.
• Help shape anti-fraud rules and policies.

Skills

SQL

Python

ML

LLMs

Data pipelines

Airflow

Product sense

Tools

ClickHouse

Databricks

Airflow

Job description

Toloka AI is building fraud-detection systems at the intersection of analytics and applied ML. You will analyze data to uncover patterns, generate hypotheses, and develop defenses—rules, models, and LLM-based approaches—deployed in production.

You’ll own end-to-end delivery from insights to monitoring, and you’ll build data pipelines (SQL, Python, Airflow), design production metrics, and contribute to anti-fraud policy shaping in a remote-friendly hybrid setup.