Security AI Engineer, Principal Member of Technical Staff

QuantumScape Corporation•San Jose, CA

14d•$155,000 - $236,000•Onsite

About The Position

QuantumScape is seeking a Principal Security AI Engineer to join their Cybersecurity Team. This team operates like a product engineering group, building platforms, agents, and automation to defend the company's intellectual property. The role involves developing an internal agentic AI platform for security and IT operations, including multi-agent workflows for incident triage, asset reconnaissance, access reviews, and audit evidence generation. The ideal candidate is excited about applying AI to complex security challenges in a fast-paced, technical environment.

Requirements

Bachelor's degree in Computer Science, Computer Engineering, or a related technical field and 10+ years of hands-on experience across cybersecurity, cloud infrastructure, or platform engineering, with demonstrated depth in cloud security, networking, and automation
Proven track record shipping production software or infrastructure — not just scripts. You have built systems that other people depend on daily
Hands-on experience building with large language models: agent frameworks, tool and function calling, retrieval, prompt engineering, and systematic evaluation
Track record of building or maturing security capabilities from initial gap identification through operationalization in complex, fast-paced environments
Experience working alongside engineering, IT, and infrastructure teams in an embedded or advisory capacity
AI & Agent Engineering: Practical experience with agent frameworks (Google ADK, LangGraph, or equivalent), LLM APIs (Anthropic Claude, Vertex AI, or comparable), Model Context Protocol (MCP), tool-use design, and agent evaluation. Ability to reason clearly about non-determinism, guardrails, and failure modes.
Programming: Strong production-quality Python (typing, testing, packaging, async). Working knowledge of a second language such as Go, TypeScript, or PowerShell.
Cloud Platforms: Deep GCP expertise — IAM, VPC and networking, GKE, Cloud Run, Secret Manager, Workload Identity Federation, Cloud Logging and Audit Logs, and Security Command Center. Working knowledge of Microsoft Azure and Entra ID.
Infrastructure & DevOps: Kubernetes, containers, Terraform, CI/CD pipelines, GitOps, secrets management, and observability tooling.
Networking: Strong grasp of TCP/IP, DNS, TLS, routing, firewall management, segmentation, private connectivity and interconnects, IDS/IPS and NDR tooling, and IT/OT boundary controls.
Detection & Response: Hands-on experience with SIEM and XDR platforms including Google SecOps (Chronicle), Microsoft Defender XDR, and Microsoft Sentinel; detection engineering, correlation logic, and response automation.
Security Engineering Breadth: Vulnerability management (Tenable or equivalent), endpoint control and application allowlisting (ThreatLocker or equivalent), identity and privileged access (SSO, FIDO2, PAM), and software supply chain security (SBOM, image scanning, signing).
Frameworks: MITRE ATT&CK, NIST CSF and NIST SP 800-207, CIS Benchmarks; familiarity with emerging AI security guidance such as the OWASP Top 10 for LLM Applications and the NIST AI Risk Management Framework.
Operating Systems: Solid experience across Linux and Windows environments.

Nice To Haves

Experience securing or operating manufacturing and OT environments
Experience with automotive supplier security assessments (TISAX / VDA ISA) or comparable regulated audit regimes
Contributions to open-source agent, MCP, or security tooling
CISSP (Certified Information Systems Security Professional) preferred
Google Professional Cloud Security Engineer, CKA/CKS, or equivalent cloud and Kubernetes certifications strongly considered
Relevant certifications such as GCIH, GCFA, GCIA, GCTI, or equivalent advanced certifications are strongly considered

Responsibilities

Design, build, and operate AI agents that perform real security work — incident triage, asset reconnaissance, access reviews, evidence collection, and reporting
Engineer the infrastructure those agents run on: Kubernetes, service networking, workload identity, secrets, and observability
Build and maintain Model Context Protocol (MCP) servers that connect agents to our security, cloud, and IT platforms with least-privilege, scoped tooling
Replace runbooks and manual toil with automated pipelines that close the loop into ticketing, SIEM, and GRC systems
Treat the agent platform as production security infrastructure — no public exposure of internal tooling, least privilege everywhere, full audit trail
Hold a high bar for correctness: an agent that is confidently wrong is worse than no agent at all
Harden cloud and network environments by identifying gaps, enforcing segmentation, and ensuring monitoring coverage
Communicate with clarity across all levels of the organization, adapting your message to your audience
Take problems all the way to resolution — not just to identification — closing loops and holding yourself and others accountable
Build and operate a multi-agent security platform (Python, Google Agent Development Kit, Anthropic Claude) deployed on GKE — including agent routing, session management, streaming interfaces, and multi-agent handoff
Design human-in-the-loop gating so any agent action with real-world consequence requires explicit approval, with full decision provenance retained
Develop and maintain MCP servers exposing scoped tools over Microsoft Defender XDR, Google SecOps (Chronicle), Entra ID, GCP, Jira/Confluence, and Microsoft 365
Build evaluation harnesses and regression suites that measure agent accuracy, tool-call correctness, and hallucination rate before anything reaches production
Own agent observability: prompt and response logging, tool-call auditing, cost and latency telemetry, and behavioral drift detection
Ship analyst-facing surfaces — a web frontend with SSO and role-based access control, plus agentic CLI workflows for hands-on investigation
Identify the highest-toil security workflows and rebuild them as automated pipelines: alert enrichment, phishing triage, vulnerability ticket routing, access certification, and audit evidence collection
Codify detection content and response actions as version-controlled, peer-reviewed, testable artifacts (detection-as-code)
Integrate automation into SIEM, ITSM, and GRC platforms so findings become tracked, closed-loop work rather than another dashboard
Define what stays human: escalation criteria, blast-radius limits, approval gates, and rollback paths for every automated action
Measure and report the outcome — analyst hours reclaimed, mean time to triage, coverage gained, and false-positive reduction
Design and harden GCP foundations: organization policy, IAM and service account hygiene, Workload Identity Federation, VPC Service Controls, Secret Manager, CMEK, and Cloud Audit Logs
Harden GKE end to end — control plane configuration, node hardening, workload identity, admission control, network policy, and supply chain integrity (image signing, Binary Authorization, SBOM)
Build infrastructure as code (Terraform) with policy-as-code guardrails so security posture is enforced at deploy time rather than discovered during an audit
Eliminate privilege escalation and lateral movement paths across projects, service accounts, and peered networks
Instrument cloud telemetry into Google SecOps and validate detection coverage across GCP, Azure, and SaaS log sources
Assess and harden network architecture across on-premises, cloud, and OT-adjacent environments, ensuring alignment with security best practices
Design and enforce segmentation and zero-trust access patterns, including perimeter and egress controls
Partner with IT and infrastructure teams on firewall policy management, private connectivity and cloud interconnects, VPN architecture, and secure remote access
Ensure agent and automation workloads follow strict networking principles: private in-cluster service discovery, controlled egress, and no public exposure of internal tooling
Evaluate and improve network visibility tooling (NDR, IDS/IPS, flow logging) and feed that telemetry into automated detection and hunting
Provide security guidance on network and platform architecture decisions, including IT/OT boundary controls
Own the security model for our own AI usage: prompt injection resistance, tool authorization boundaries, token scoping and lifetime, and data classification controls governing what agents are permitted to read
Define and operationalize GenAI governance — approved tooling, handling rules for classified intellectual property, and monitoring of AI application usage across the company
Review third-party AI integrations, connectors, and MCP servers, and build the controls that let the business adopt them safely rather than blocking them outright
Track the evolving AI threat landscape and translate it into concrete detections, guardrails, and architecture requirements
Act as a security advisor to engineering, platform, and IT teams, translating complex security requirements into practical, actionable guidance
Influence technology decisions, architecture reviews, and vendor assessments from a security lens
Communicate risk and program progress clearly to both technical and non-technical stakeholders, including senior leadership
Foster a culture of security ownership and automation-first thinking across the organization