AI Security Engineer

Description

The company builds browser security for both human and AI agent workforces, stopping threats across web pages, documents and email before they reach the user, and now extends that protection to the AI agents working alongside employees.

This is a Data & AI / ML Evaluation Engineer role that owns data collection, mining, curation and continuous evaluation for the company's AI security products. It is a data and ML software engineering role focused on the data infrastructure used to train its in-house models: Prompt Injection SLMs, Shadow-AI detection, Malicious Macro classifiers, PII detection, content classification and more.

You will own these outcomes: automated, continuous creation and curation of training, evaluation and false-positive datasets across web, file and macro threats; standardized, reproducible benchmarks that guide model changes before deployment; uptime and ingestion throughput of web and file data pipelines; the size, diversity and labeling accuracy of threat datasets for PI, Shadow-AI, PII and macros; how quickly live false-positive and false-negative feedback reaches training and evaluation; and accuracy and benchmark coverage of historical versus newly trained SLMs.

Responsibilities

• Design and run automated pipelines that collect, mine and scrape threat samples from surrogate traffic and web sources.
• Build processing, annotation and curation workflows, and train and retrain models such as Prompt Injection SLMs, Shadow-AI detectors, Malicious Macro classifiers and PII detectors.
• Build ways to capture false-positive samples from live user and DDR flows and feed them into retraining pipelines.
• Build automated benchmarking and evaluation harnesses that test detection quality, runtime performance and accuracy against baselines such as Vertex AI, on a continuous basis.
• Build automated training pipelines for the different models.
• Work with research and software engineering teams to turn raw detection telemetry into steady, actionable model improvements.

Requirements

• Have 2, 4 years of experience in software engineering, data engineering, or automated QA and evaluation.
• Be strongly proficient in Python.
• Have experience training models and working with machine learning and AI algorithms.
• Have built data pipelines, web scrapers or automated data processing systems.
• Understand data structures, storage formats and database configuration.
• Work methodically on dataset labeling, validation, bug logging and metric tracking.
• Nice to have: experience with ML evaluation frameworks, dataset curation or annotation tools.
• Nice to have: familiarity with text extraction, NLP preprocessing or document parsing.
• Nice to have: understanding of cybersecurity concepts such as prompt injection, malicious macros and DLP or PII.
• Nice to have: basic experience with cloud storage, object stores and Docker.
• Nice to have: experience with AWS, GCP, Terraform, Kubernetes and Linux.
• Nice to have: C++ and SQL would be an advantage.