ML Scientist – Security & Fraud (Hybrid)

OneSpan Inc.; Barcelona; Hybrid; EUR 90.000 - 120.000; Vollzeit

Zusammenfassung

OneSpan Inc. is seeking a senior ML researcher to tackle adversarial security challenges. You will work alongside the Security team to translate attack techniques into data problems, analyze telemetry from ClickHouse and S3, and prototype scalable detection models.

You will define metrics for detection performance and validate prototypes, aiming for production integration with engineering. A strong research background in security ML and experience with large telemetry are required.

Qualifikationen

• 5+ years researching and developing ML models in security, fraud, or adversarial domains.
• Strong anomaly detection and working with scarce labels.
• Experience with large-scale telemetry or time-series data.

Aufgaben

• Work with Security to understand attack techniques against mobile apps and translate them into data problems.
• Analyze telemetry in ClickHouse and S3 to find signals, patterns and anomalies.
• Design, prototype, and evaluate detection and prevention models (anomaly, classification with scarce labels).
• Define success metrics, false-positive rates, and evaluation methods without full ground truth.
• Take validated prototypes toward production with engineering, building features and labeling pipelines.
• Stay current with ML security research and bring actionable ideas.

Kenntnisse

Anomaly detection

Unsupervised learning

Semi-supervised learning

Python (pandas/polars)

SQL

PyTorch

Telemetry data

Ausbildung

PhD or equivalent experience

Tools

AWS S3

ClickHouse

MLOps tooling

Jobbeschreibung

OneSpan Inc. is seeking a senior ML researcher to tackle adversarial security challenges. You will work alongside the Security team to translate attack techniques into data problems, analyze telemetry from ClickHouse and S3, and prototype scalable detection models.

You will define metrics for detection performance and validate prototypes, aiming for production integration with engineering. A strong research background in security ML and experience with large telemetry are required.