EmergingAI control, monitoring and containmentPart of The agent workforce

Agent Containment Engineer

Also called AI Sandboxing Engineer

An Agent Containment Engineer builds the boundaries that keep AI agents from doing harm, even if the model misbehaves or is attacked. They matter because agents with real access are only as safe as the walls around them.

What they do

They design and build the environments agents run in. That means isolating code execution, limiting what files, networks and tools an agent can touch, and controlling what leaves the environment. They apply least privilege, so an agent gets only the access a task needs. They write tests that try to break out, and tests that check policy is enforced. They work with evaluation and red teams who attack the sandbox with permission. They add monitors that flag odd behaviour at the boundary. When something gets through, they find the cause and close it.

What a strong hire delivers

In the first months, a strong hire maps what your agents can currently reach and where the real risks sit. They ship a hardened sandbox for one live agent workload, with egress rules and permission limits in place. They leave behind a suite of escape and policy tests that runs on every change, plus a clear record of what the boundary does and does not guarantee.

How you can tell they are good

Ask them to walk through a boundary they built and how they tried to break it. Strong people name specific escape routes, such as network side channels or over-broad credentials, and explain what they fixed. They are honest about residual risk. Ask how they would test a sandbox against an agent that is actively hostile. Be wary of anyone who talks only about policy documents or says a sandbox is simply secure.

Why it is hard to fill

Almost nobody holds this title yet. The work usually sits inside security, infrastructure or safety teams under other names. The skill mix is odd: deep isolation knowledge plus an understanding of how models and agents behave. Security engineers often lack the model side. Safety researchers often lack hardening experience. Candidates who have both are few, and most are not looking.

Where to find them

Look first at security engineers who have worked on isolation, virtualisation, container hardening or egress control. Then look at people from formal methods, and at AI safety evaluation teams who have built or attacked agent environments. They sit in frontier labs, AI security start-ups, cloud infrastructure teams and research groups focused on AI control. Many carry titles like security engineer or evaluation engineer.

Indicative pay

Not enough published evidence yet to give a fair range for this role, which is common for work this new. We benchmark it for you as part of a search.

Interview questions for an Agent Containment Engineer

  1. Describe a sandbox or isolation boundary you built. How did you test that it held?

    A strong answer names the threat model, the specific controls and the tests used, including attempts to escape. Look for honest limits. Worry if they describe only configuration, with no adversarial testing or evidence.

  2. How would you decide what network access an agent should have for a task?

    They start from least privilege, define the task needs, default to deny and use allowlists with logging. They discuss exfiltration paths. Worry if they default to open access and rely on monitoring alone.

4 more questions, with what a strong answer covers, are in the Role Kit.

Job description you can post

We are building agents that act in real systems, and we need to be sure they stay within set limits even if the model behaves badly or is attacked. We are hiring an Agent Containment Engineer to design, harden and test the environments our agents run in.

What you will do

  • Design and build isolated execution environments for agents
  • Define and enforce egress controls and least-privilege access
  • Write escape and policy tests and run them on every change
  • Work with evaluation and red teams on authorised adversarial testing
  • Build monitoring that flags boundary probing and policy breaches
  • Investigate incidents and close the gaps they reveal
  • Document what each boundary does and does not guarantee

What you will bring

  • Demonstrated work hardening isolation boundaries, such as containers, virtual machines or sandboxes
  • Hands-on experience with egress control and permission design
  • Experience attacking or testing systems for escape and privilege problems
  • Clear thinking about threat models and residual risk
  • Ability to write production-quality code and tests
  • Comfort working with researchers and explaining risk plainly

Nice to have

  • Experience with formal verification or formal methods
  • Background in AI safety evaluation or AI control
  • Familiarity with agent frameworks and how agents use tools

Free to use and adapt. If it saves you time, a link back to this page is appreciated.

Skills that matter

  • AI controlAdvancedLimit harm from potentially untrusted AI behaviour
  • Security engineeringAdvancedProtect systems, identities and sensitive assets
  • Model evaluationWorkingMeasure capability and failure modes credibly
  • Formal verificationWorkingProve stated properties under explicit assumptions

Other roles in aI control, monitoring and containment

EmergingAI Control ResearcherAn AI Control Researcher designs the safeguards that limit what a model can do if it turns out to be unreliable or hostile. They keep risk bounded…EmergingAI Monitor Evaluation EngineerAn AI Monitor Evaluation Engineer tests whether the systems that watch model behaviour actually catch dangerous or unauthorised actions. It matters…EmergingAI Shutdown and Recovery EngineerAn AI Shutdown and Recovery Engineer builds and tests the ways to stop an AI system mid-action and bring it back to a known good state. It matters…EmergingAI Control EngineerAn AI Control Engineer builds the monitoring, sandboxing and containment that let a company use powerful models safely, even if a model is…

Related roles in other families

EmergingAI Formal Verification EngineerFormal methods and reliable reasoningAn AI Formal Verification Engineer writes precise specifications for AI systems and their outputs, then proves they hold or finds counterexamples…Hiring nowAutomated Theorem Proving ResearcherFormal methods and reliable reasoningAn Automated Theorem Proving Researcher builds and improves methods that find or check mathematical and logical proofs by machine. The company gets…Hiring nowProgram Synthesis ResearcherFormal methods and reliable reasoningA Program Synthesis Researcher builds systems that generate programmes which meet a testable or formal specification. They give the company code…EmergingAI-Assisted Verification SpecialistFormal methods and reliable reasoningAn AI-Assisted Verification Specialist uses AI tools to speed up proof and specification work, then checks every result independently. They help a…

How we wrote this

Role definitions are our own editorial work, checked against occupational classifications, standards, research and employer pages. Titles vary by company, and this is a description of the work, not a job advert.

← AI control, monitoring and containmentAll new job families

Start a search

Hiring an Agent Containment Engineer?

Tell us the work. We find the people already doing it, whatever their job title says, and you meet them already interviewed for your role.