Software Engineer, Security

Notion Notion · Enterprise · San Francisco, CA · Engineering

Experienced security engineer to own cross-cutting programs at the intersection of product, infrastructure, and AI. Responsibilities include modernizing authentication, building and operating AI safety guardrail infrastructure (prompt-injection protections, provenance systems), and advancing the authorization platform. The role requires shipping security-critical infrastructure and experience with AI/LLM security protections or a clear ability to ramp quickly.

What you'd actually do

  1. Modernize and migrate authentication across Notion’s product surfaces (SAML/OIDC, OAuth flows, session semantics, passkeys, CSP, redirect handling), landing multi-quarter changes with clear rollout plans and minimal customer disruption.
  2. Build and operate Notion’s AI safety guardrail stack, including prompt-injection protections (vendor evaluation, deployment model decisions, integration with agents) and an external-source provenance system for AI-generated content across Mail, Calendar, and MCP.
  3. Advance our authorization platform direction by driving crisp architectural trade-offs (e.g., SpiceDB vs. Macaroons) and shipping reusable primitives that product teams can adopt without bespoke security work.
  4. By day 90: own one P0 security program end-to-end—RFC, rollout plan, partner alignment, execution, and measurable risk reduction—plus ship one piece of AI leverage (e.g., an internal security agent for triage/verification/continuous checks) that improves correctness and reduces time-to-resolution.
  5. By end of year 1: raise the bar on security engineering craft by setting clearer standards for secure primitives (auth/authz, provenance, domain posture), improving adoption paths for partner teams, and reducing recurring classes of vulnerabilities through better systems—not heroics.

Skills

Required

  • Demonstrated ability to ship security-critical infrastructure in production systems (identity/authentication, authorization, platform primitives), including migrations that affect customers and require careful rollout and backwards compatibility.
  • Strong judgment navigating ambiguous trade-offs (security vs. product velocity, correctness vs. ergonomics, centralized platforms vs. local autonomy), with a track record of writing clear RFCs and aligning cross-functional stakeholders.
  • Experience building or operating AI/LLM security protections (e.g., prompt injection, tool/data provenance, policy enforcement) or a clear ability to ramp quickly and lead in an emerging domain.
  • High agency and systems mindset: you proactively find the real constraint, unblock partner teams, and build primitives that compound across the org (not one-off fixes).
  • Comfort mentoring and multiplying others—through intern/project ownership, enablement sessions, and pragmatic security guidance that engineers actually adopt.

What the JD emphasized

  • AI guardrail infrastructure
  • prompt-injection protections
  • external-source provenance system for AI-generated content
  • AI leverage (e.g., an internal security agent for triage/verification/continuous checks)
  • AI/LLM security protections

Other signals

  • AI agent safety
  • AI guardrail infrastructure
  • prompt-injection protections
  • external-source provenance system for AI-generated content
  • internal security agent for triage/verification/continuous checks