Real-Time Fraud ML Engineer (Remote-First)

Sardine; United States; À distance; USD 140 000 - 220 000; Plein temps

Avantages offerts par ce poste

Remote-first culture

Health insurance

MacBook Pro

Stipends for meals

Home office stipend

Learning stipend

Annual health & wellness stipend

401(k) / RRSP matching

Flexible PTO

Résumé du poste

Sardine is hiring a Machine Learning Engineer to own the real-time fraud-detection platform. You will build and operate the model-serving stack, ensure ultra-low latency, and deploy models with a focus on reliability, observability, and security across a remote-first, US/Canada-friendly environment.

You'll work with Python and Go, leverage Kubernetes on GCP, and contribute tooling for self-serve ML workflows while maintaining strict compliance in a regulated context.

Qualifications

• Experience building, not just using, model serving infrastructure.
• Production ownership of ML systems: paged when something broke, root cause, fix preventing recurrence.
• Strong Python and solid software engineering fundamentals (testing, code review, CI/CD).
• Comfort with Kubernetes, containers and Google Cloud, plus infrastructure-as-code.

Responsabilités

• Build and own the model serving infrastructure and real-time inference with latency budgets.
• Build deployment path for data scientists to ship models, including BYO-model for clients.
• Own production models: monitoring, drift detection, retraining, incident response, on-call.
• Create pipelines turning raw signals into production-ready features.
• Collaborate across Python and Go for fast inference in the request path.
• Develop tooling for engineers/data scientists to use; balance self-serve vs controlled access.
• Champion testing, observability, security and compliance in a regulated environment.

Connaissances

Model serving infra

Production ownership

Python

Kubernetes

GCP

CI/CD

Debugging ML models

Self-serve tooling

Outils

Docker

Description du poste

Sardine is hiring a Machine Learning Engineer to own the real-time fraud-detection platform. You will build and operate the model-serving stack, ensure ultra-low latency, and deploy models with a focus on reliability, observability, and security across a remote-first, US/Canada-friendly environment.

You'll work with Python and Go, leverage Kubernetes on GCP, and contribute tooling for self-serve ML workflows while maintaining strict compliance in a regulated context.