Staff Security Engineer - Product Security

Zipline•South San Francisco, CA

4h•Hybrid

About The Position

Zipline is the world’s largest and most experienced drone delivery service, on a mission to serve all humans equally by ensuring access to food, medicine and essential goods anytime, anywhere. They design, build, and operate the world’s largest autonomous logistics system, making millions of deliveries globally. The company operates on four continents, making a delivery every 30 seconds, including blood, vaccines, medical supplies, food, and retail products. Zipline's system strengthens supply chains, reduces congestion, and gives people time back, with over 140 million commercial autonomous miles safely flown. The software team is building scalable solutions to expand rapidly and empower distribution centers. Zipline’s security challenges involve real-world autonomy, robotics, global operations, cloud software, and regulated/health-adjacent workflows, presenting a large attack surface. The ideal candidate will work in a startup environment, collaborate across engineering disciplines, and join a small, high-ownership security team with significant influence. The role also involves helping Zipline safely leverage LLM copilots and agentic tools while containing risks related to secrets, sensitive data, and privileged actions, aligning with frameworks like NIST AI Risk Management Framework and OWASP Top 10 for LLM Applications.

Requirements

8+ years of experience designing, building, and operating security controls for large-scale production systems (application, cloud, and infrastructure security).
Strong security engineering chops with evidence you can reduce risk in production systems (not just talk about it).
Hands-on ability to write and ship code/tools in Python, Go, or similar (you’re expected to build, not just review).
Practical experience securing microservice architectures and modern cloud stacks (containers/Kubernetes, IAM, CI/CD, secrets, logging).
Comfort operating as a technical leader without authority: you can persuade, teach, and unblock - not police.
A skeptical mindset: you naturally ask “what’s the failure mode?” and “how will this be abused?” before shipping changes.
Familiarity with the security failure modes of LLM-enabled systems (or the willingness to learn fast), including risks called out by OWASP such as prompt injection, insecure output handling, insecure plugin design, and excessive agency.

Nice To Haves

Experience spanning multiple engineering domains (web app + cloud infra + embedded/robotics/autonomy).
Experience building developer-friendly security platforms (internal libraries, paved roads, CI integrations, Public Key Infrastructure).
Track record of being an effective security “evangelist” (i.e., enabling good behavior with good tools and defaults, not fear).
Experience designing guardrails for internal AI/agent usage (policy + technical controls + auditing), especially in environments where safety and reliability are non-negotiable.
Deep understanding of distributed systems and how failures actually happen (partial outages, weird retries, cascading dependencies, misconfigurations, permissions drift).

Responsibilities

Own security outcomes for critical parts of Zipline’s application and cloud ecosystem (not by writing policy docs that no one reads, but by shipping controls and enabling teams).
Partner with engineering teams on secure architecture, threat modeling, and design reviews for services that must be correct, reliable, and defensible under real-world operational pressure.
Help us build and scale a pragmatic secure SDLC – CI/CD hardening, dependency/supply-chain controls, secrets management, and code review patterns that don’t slow teams down.
Improve cloud security posture end-to-end: IAM and least privilege, network/service-to-service trust, key management, logging/telemetry, runtime detection, and incident-ready auditability.
Drive vulnerability management that actually closes risk: triage, exploitability analysis, remediation partnerships, and verification.
Help build and exercise incident response: playbooks, tabletop exercises, logging requirements, and “know it happened / know what changed” operational discipline.
Support data classification and access control models aligned to how Zipline operates (including partner/customer interfaces and global operations).
Support external penetration tests and turn results into durable improvements, not whack‑a‑mole patches.
Contribute to security compliance efforts (e.g., SOC 2 / ISO 27001) in a way that strengthens engineering
Secure AI-assisted and agentic engineering workflows (this is explicitly part of the job):
define safe patterns for copilots/LLM tools used in development and ops
implement guardrails for sensitive data exposure and output handling
prevent “agentic overreach” (over‑privileged tools, unsafe tool-calling, silent action-taking)
build monitoring/auditing around AI tool use where it matters