Hone builds AI that owns outcomes, not tasks: agentic systems that work towards goals over weeks and months, deploy themselves, and improve over time.
About Hone
Models keep getting smarter, but outside of coding most organizations can't absorb that intelligence, and the gap between model capability and the value derived from it keeps widening. The reasons are many, from human incentives to org design, but at the core we think it's a technology problem. Hone builds engines: AI systems that you hand a goal, like "keep the recruiting pipeline healthy." An engine builds itself, maintains its own agents and code, pulls in people when it needs judgment, and learns from every correction. Leading AI companies and global enterprises already run engines across funnel optimization, procurement, deal sourcing, asset management, financial risk, and product and customer delivery. We measure ourselves by the real-world economic impact our engines create, not benchmark scores. We're ex-founders and early engineers from Cognition, Mercor, Ramp, Jane Street, Decagon, Brex, and Harvey, backed by Benchmark and Index Ventures.
The Role
Engines act inside a company's most sensitive systems for weeks or months, with real permissions and on their own schedule. That makes security a new kind of problem: any input can be an attack, the actor is an agent, and the blast radius is vast. You'll own security across Hone: the isolation and identity model every engine runs under, the controls that let enterprises trust engines with their systems and data, and the defenses against agent-specific threats.
What You'll Work On
Isolation and network sandboxing for long-running agents: default-deny egress, scoped access, tenant isolation.
Identity and credentials for engines: per-engine service identities, short-lived scoped credentials, and a broker so agents never hold raw secrets.
Defenses against prompt injection, data exfiltration and tool misuse across untrusted inputs.
Permissions and audit: RBAC, approval gates, and a full record of every action an engine takes and why.
Security architecture for single-tenant and customer-VPC deployments.
Internal AI tooling for detection, review and response, including agents that red-team our own engines.
Requirements
2+ years of security engineering experience, or equivalent depth from research, CTFs, open source or your own projects
Time at a fast-growing startup
You've designed and built security systems from scratch, not just audited them
You've secured production systems under real load and handled real incidents
Strong fundamentals: cloud security, identity and access, networking, isolation, application security
You use coding agents every day and still understand every line you ship
You break and build things for fun because you genuinely care about how systems fail
Nice to Have
Agent or LLM security experience: prompt injection, sandbox escapes, red-teaming models or agents
You've taken an enterprise product through security reviews and audits. A CTF background is a plus.
How We Hire
Intro with a co-founder.
Technical screen.
Security deep dive.
Two-day work trial in San Francisco.
References.
Offer.
Logistics
Location: San Francisco, in person, Sunday to Friday.
Compensation: Competitive salary + meaningful equity.
Visa: We sponsor visas and support you through the immigration process.
Benefits: Health and dental insurance, maternity and parental support, and relocation support.
