Sardine is hiring a remote Machine Learning Engineer!
Sardine runs an agentic risk platform that helps banks, fintechs, and SaaS companies stop fraud and financial crime in real time, drawing on a large consortium of profiled devices, consumers, and businesses. The team is remote-first across the US and Canada, hires self-directed people, and judges performance by output rather than hours. This Machine Learning Engineer role sits on the platform side of the data science team: client-specific models are built by data scientists, and you own the infrastructure they are deployed onto and the low-latency scoring path they run in. Every session has to be scored within roughly 250 milliseconds from hundreds of device and behavioural signals, which drives decisions about how features are computed and served, how models ship and roll back, and how quickly degradation is detected. Expect deep production ownership, including an on-call rotation, with about a fifth of the time spent building models directly — more if desired.
Key responsibilities
- Own model serving infrastructure, real-time inference, and feature retrieval within a strict latency budget
- Build and improve the self-serve deployment path data scientists use to ship models, including support for clients running their own models
- Operate models in production: monitoring, drift detection, retraining, incident response, and on-call participation
- Develop and tune pipelines that convert raw device and behavioural signals into production-ready features
- Work across Python and the Go backend to keep inference fast inside the live request path
- Build models directly for roughly 20% of the role, with room to do more
- Advocate for testing, observability, security, and compliance in a regulated setting
Key requirements
- Hands-on experience building model serving infrastructure rather than only consuming it
- Prior production ownership of ML systems, including diagnosing failures and preventing recurrence
- Strong Python plus solid engineering discipline: testing, code review, CI/CD
- Comfort with Kubernetes, containers, a major cloud provider (GCP preferred), and infrastructure-as-code
- Enough model understanding to distinguish data issues, feature pipeline issues, and model issues when quality drops
- Experience shipping tooling that other engineers or data scientists adopt, with judgement about what should be self-serve
- Bonus: fraud, risk, or cybersecurity domain knowledge; browser API and high-entropy data collection familiarity; experience applying frontier LLMs to automation
- Seniority: Not stated
- Work mode: Remote (US or Canada)
- Key skills: Python, Go, Kubernetes, Docker, GCP, Infrastructure as Code, CI/CD, Model serving, Real-time inference, Feature pipelines, Observability, Drift detection
Want more jobs?
Join our mailing list to receive updates about new data science and Machine Learning Engineer jobs.
