ML Ops Engineer

The role

You will own the platform that takes fraud, risk, and payments models from notebook to production. You will build pipelines, serving infrastructure, and monitoring to score transactions in real time.

What you will do

• Build and operate the inference layer for real-time model serving.
• Design safe fallbacks for model or feature unavailability.
• Own end to end training, validation, and deployment pipelines.
• Build reproducible feature pipelines from transaction, device, and chargeback data.
• Manage model versioning, registries, shadow mode, and rollouts.
• Instrument models to track latency, score distributions, and feature drift.
• Integrate and evaluate external fraud and risk vendors.
• Power data and scoring for merchant-facing risk products.

Required qualifications

• 4+ years in MLOps, ML infrastructure, or ML platform engineering
• Experience running models in production with real latency and uptime requirements
• Strong Python
• Solid production engineering habits
• Comfortable reading and writing TypeScript/Node
• Hands-on experience with ML pipeline and serving tooling (MLflow, Kubeflow, SageMaker, Vertex, BentoML, Feast, or similar)
• Solid infrastructure skills: containers, Kubernetes, CI/CD, and at least one major cloud provider
• Ability to build reliable feature pipelines over operational, warehouse, and streaming data
• Understanding of training/serving skew
• Experience monitoring models in production: drift, calibration, and tying model metrics to business outcomes
• Familiarity with fraud, risk, or credit modeling
• Ability to explain model behavior to non-ML stakeholders like banks and auditors

Nice to have

• Experience with real-time inference in a transaction or decisioning path

About the company

Coinflow is a payment service provider serving marketplaces, fintechs, remittance providers, gaming platforms, and ecommerce merchants worldwide. The company is headquartered in Chicago, IL.