Machine Learning Engineer III, Shield

Box•Redwood City, CA

4d•$175,500 - $219,500•Hybrid

About The Position

Box is hiring a Machine Learning Engineer for its Shield team, which builds security capabilities that protect enterprise content and help keep Box’s Intelligent Content Management platform secure. The role focuses on building and scaling machine-learning systems to help protect millions of users and billions of files from ransomware, data exfiltration, insider threats, and emerging AI-based attacks. As part of a fast-paced backend and core team, the engineer will collaborate across Box to deliver security capabilities for high-demand use cases. The position designs, trains, and deploys threat-detection models; owns end-to-end machine-learning data pipelines; develops features for real-time and batch anomaly detection; and deploys, monitors, and improves production models serving enterprise customers. The engineer partners with Platform, Application Engineering, and Product teams to turn security requirements into machine-learning solutions and participates in an on-call rotation. Box expects employees to work from their assigned office at least three days per week. The role is eligible for equity and benefits. The advertised Redwood City base salary range is $175,500–$219,500 USD.

Requirements

At least 3 years of experience in applied machine learning.
Strong programming skills in Python.
Experience deploying and maintaining machine-learning models serving real traffic.
Experience with GCP (Vertex AI, BigQuery, Dataflow) or equivalent platforms such as AWS SageMaker or Azure ML.
Deep understanding of feature engineering, model evaluation, and MLOps.
Strong communication skills, including the ability to explain complex machine-learning concepts to non-technical stakeholders.
Work from the assigned office at least 3 days per week.

Nice To Haves

Experience in security or threat detection, anomaly detection, or fraud detection.
Experience with sequential data, anomaly detection, or behavioral modeling, including time-series forecasting, LSTMs, Transformers, or similar methods.
Knowledge of LLMs and AI safety, including prompt injection, guardrails, and red-teaming.
Experience with streaming or real-time machine-learning systems.
Familiarity with Java for service integrations.
Publications or contributions in machine-learning security.

Responsibilities

Design, train, and deploy machine-learning models for ransomware detection, suspicious-session identification, and user behavior analytics.
Own end-to-end machine-learning pipelines that process high-volume security event streams using Apache Spark, GCP Dataflow, and Vertex AI.
Create and maintain feature stores that power real-time and batch anomaly-detection systems.
Deploy, monitor, and iterate on machine-learning models in production for enterprise customers at scale.
Partner with Platform, Application Engineering, and Product teams to translate security requirements into machine-learning solutions.
Participate in the on-call rotation and be available while on call to respond to and triage issues.