Senior ML Engineer, AI Safeguards & Production Security

ServiceNow; Mountain View (CA); On-site; USD 190,000 - 270,000; Full time

Job summary

ServiceNow in Mountain View, CA seeks a Senior AI/ML Engineer to design, build and ship safeguards for large language models, including guardrail models and jailbreak detection, working with security and product teams to reduce risk while enabling practical AI features. Requirements include 8+ years in production software, 3+ years in applied ML, excellent Python skills, and strong model evaluation.

You will optimize latency and cost, and define configurable safeguard policies for enterprise

Qualifications

• 8+ years building production software, with 3+ years of applied ML.
• Experience safeguarding LLMs: guardrail models, content classifiers, or safety evals.
• Strong model evaluation for agent trajectories and text classification.
• Understanding bypass methods like indirect prompt injection and multi-turn jailbreaks.
• Excellent Python experience.

Responsibilities

• Design and ship safeguards in the agent harness for prompt injection and jailbreaks.
• Build evals measuring safeguard efficacy from reports and production data.
• Collaborate with Security to turn findings into cases and mitigations.
• Select and tune detection models for commercial and government environments.
• Balance quality, latency, and cost for in-path controls.
• Create configurable safeguard policies for customers.
• Monitor production safeguards and fix regressions.

Skills

Production software

Applied ML

Python

Guardrails

Model evaluation

Prompt injection

LLM safety

Data quality

Job description

ServiceNow in Mountain View, CA seeks a Senior AI/ML Engineer to design, build and ship safeguards for large language models, including guardrail models and jailbreak detection, working with security and product teams to reduce risk while enabling practical AI features. Requirements include 8+ years in production software, 3+ years in applied ML, excellent Python skills, and strong model evaluation.

You will optimize latency and cost, and define configurable safeguard policies for enterprise