Production ML Engineer - Systems & Reliability

Sprinter Health; San Francisco (CA); Hybrid; USD 150,000 - 200,000; Full time

Benefits offered by this job

Medical, dental, and vision plans 100%

Flexible PTO

401(k) with match

Free daily lunch in-office

Learning stipend

Job summary

Sprinter Health in San Francisco is seeking an ML Engineer to build production systems that train, deploy, monitor, retrain, and serve machine-learning models reliably. You will bridge software engineering, data engineering, and modeling to make ML work in production.

You’ll construct training and inference pipelines, serve predictions via APIs and batch jobs, and implement monitoring to catch drift before it affects patients or partners. You think in systems, not notebooks.

Qualifications

• Strong Python and software-engineering fundamentals.
• Experience with ML frameworks, data pipelines, and model serving.
• Experience taking models from prototype to reliable production.
• Cloud infrastructure, containers, CI/CD, and orchestration.
• Monitoring, observability, reproducibility, and versioning across data, features, and models.
• Security and privacy controls for sensitive data.

Responsibilities

• Build and harden training pipelines.
• Package models for deployment.
• Serve predictions through APIs or batch jobs with reliability in mind.
• Maintain feature pipelines and keep features fresh and correct.
• Monitor drift, data quality, latency, cost, and performance.
• Automate retraining and validation, and design safe rollback.

Skills

Python

ML frameworks

Data pipelines

Model serving

Cloud infrastructure

CI/CD

Containers

Observability

Security & privacy

Tools

Kubernetes

Job description

Sprinter Health in San Francisco is seeking an ML Engineer to build production systems that train, deploy, monitor, retrain, and serve machine-learning models reliably. You will bridge software engineering, data engineering, and modeling to make ML work in production.

You’ll construct training and inference pipelines, serve predictions via APIs and batch jobs, and implement monitoring to catch drift before it affects patients or partners. You think in systems, not notebooks.