Member of Technical Staff – Security

Hone•San Francisco, CA

21h•Onsite

About The Position

Hone builds agentic AI systems that work toward goals over weeks and months, deploy themselves, and improve over time. These engines operate within customers’ sensitive systems, with real permissions and on their own schedules. Hone describes security in this setting as a new kind of problem: any input can be an attack, the actor is an agent, and the potential blast radius is vast. The Member of Technical Staff – Security will own security across Hone, including the isolation and identity model used by every engine, the controls that enable enterprises to trust engines with their systems and data, and defenses against threats specific to agents. The role is based in San Francisco and is in person, Sunday through Friday. Hone offers competitive salary and meaningful equity, sponsors visas, and provides health and dental insurance, maternity and parental support, and relocation support.

Requirements

At least 2 years of security engineering experience, or equivalent depth from research, CTFs, open source, or personal projects.
Experience working at a fast-growing startup.
Experience designing and building security systems from scratch, rather than only auditing them.
Experience securing production systems under real load and handling real incidents.
Strong fundamentals in cloud security, identity and access, networking, isolation, and application security.
Use coding agents every day while understanding every line of code shipped.
Genuine interest in breaking and building systems and understanding how they fail.

Nice To Haves

Agent or LLM security experience, such as prompt injection, sandbox escapes, or red-teaming models or agents.
Experience taking an enterprise product through security reviews and audits.
A CTF background.

Responsibilities

Build isolation and network sandboxing for long-running agents, including default-deny egress, scoped access, and tenant isolation.
Design identity and credential systems for engines, including per-engine service identities, short-lived scoped credentials, and a broker that prevents agents from holding raw secrets.
Develop defenses against prompt injection, data exfiltration, and tool misuse across untrusted inputs.
Build permissions and auditing systems, including role-based access control, approval gates, and a full record of each engine’s actions and rationale.
Develop security architecture for single-tenant and customer-VPC deployments.
Build internal AI tooling for detection, review, and response, including agents that red-team Hone’s engines.