Production ML Engineer: Real-Time Fraud & Scale

Sift; San Francisco (CA); On-site; USD 160,000 - 230,000; Full time

Job summary

Sift is seeking a Machine Learning Engineer to bridge data science with large-scale distributed systems. You will build end-to-end pipelines that extract signals, train models per merchant, and serve predictions at production scale with low latency.

You will work on an automated ML ecosystem that recalibrates models based on streaming telemetry data, coordinating with Core Infrastructure, Product, and Data Science teams to craft robust algorithms for fraud detection.

Qualifications

• 4+ years of professional experience building and deploying large-scale machine learning models into high-traffic production environments.
• Strong programming foundations in Java/Scala and Python for production backend and data analysis.
• Hands-on experience with distributed data processing frameworks and NoSQL stores.

Responsibilities

• Design, build, and deploy online ML models to catch evolving fraud signals in real time.
• Engineer high-frequency time-series features from massive event streams for low-latency inference.
• Maintain and enhance automated ML training and deployment pipelines (CI/CD).
• Write high-performance code to minimize scoring latency across distributed systems.
• Collaborate with Infrastructure, Product, and Data Science teams to translate fraud patterns into algorithms.

Skills

ML model deployment

Java/Scala

Python

Distributed systems

Big data

Tools

Databricks

Apache Spark

Apache Flink

Hadoop

Bigtable

Docker

Kubernetes

Kafka

Claude Code

Job description

Sift is seeking a Machine Learning Engineer to bridge data science with large-scale distributed systems. You will build end-to-end pipelines that extract signals, train models per merchant, and serve predictions at production scale with low latency.

You will work on an automated ML ecosystem that recalibrates models based on streaming telemetry data, coordinating with Core Infrastructure, Product, and Data Science teams to craft robust algorithms for fraud detection.