ML Engineer I: End-to-End Email Security

Abnormal Security Corp.; United States; Remote; USD 55,000 - 86,000; Full time

Job summary

Abnormal AI is seeking a Machine Learning Engineer - I to join the Misdirected Email Detection team. You will own end-to-end ML systems—from data wrangling and feature engineering to model training, deployment, and monitoring—delivering practical solutions at scale.

The role emphasizes production-ready experiments, evidence-based improvements, and close collaboration with product, engineering, and security teams to reduce misdirected emails in real-world environments.

Qualifications

• BS degree in Computer Science, Machine Learning, Artificial Intelligence, Information Systems, or a related field.
• 1+ years building and operating applied ML features in production systems.
• Experience end-to-end ML systems: data wrangling, feature eng, model training, evaluation, deployment, monitoring.
• Experience implementing and reasoning about algorithms, features, signals, and numerical computing.
• Experience with online vs offline pipelines, labeling workflows, and safe scalable deployments.

Responsibilities

• Own the full ML lifecycle for Misdirected Email from data wrangling to deployment and monitoring.
• Run experiments and evaluations (offline metrics, online AB tests, post-launch monitoring).
• Set thresholds and conduct error analysis to prevent regressions.
• Communicate across time zones and maintain technical docs.
• Participate in on-call rotation focusing on detection efficacy and realtime scoring.

Skills

End-to-end ML

Applied ML

Data wrangling

Feature engineering

Production deployment

Education

BS degree

Tools

Python

Go

AWS

Spark

Databricks

Job description

Abnormal AI is seeking a Machine Learning Engineer - I to join the Misdirected Email Detection team. You will own end-to-end ML systems—from data wrangling and feature engineering to model training, deployment, and monitoring—delivering practical solutions at scale.

The role emphasizes production-ready experiments, evidence-based improvements, and close collaboration with product, engineering, and security teams to reduce misdirected emails in real-world environments.