Chief Technology Officer

The bottom line: we’re building the kill switch layer for fully autonomous AI, and we need someone to own the entire technical surface. This is the technical co-founder seat.

CTRL is the out-of-band control plane for autonomous AI. Cross-layer behavioral analysis and enforcement, application to bare metal, under one control ontology. One policy reasoning coherently from an API call to a syscall to a power rail. Sub-100ms, at machine speed and scale. Not a governance dashboard. Not a compliance checkbox. The layer that holds the last word when an autonomous system starts doing things nobody signed off on.

Why this matters

Everyone is racing to ship more autonomy. Models, agents, robots, autonomous platforms, all of it moving faster and touching more. The people building autonomy capture the upside. The people running it inherit the downside: money moves, credentials spread, networks fan out, and in physical systems, equipment energizes and actuators engage. At machine speed, by the time a human notices, the blast radius is already set.
Every other serious domain solved this long ago. Industry uses safety instrumented systems. Finance uses independent risk. Critical infrastructure uses out-of-band failsafes. The rule is simple: the system doing the work is never the final authority over its own failure. AI mostly ignores that rule. We’re fixing it.
The rogue-AI story usually shows up with a red-eyed robot on the poster. The real version is quieter and more serious: a system operating beyond its intended bounds, faster than any human can track, with no independent way to take the keys back. We build the layer that takes them back. Cloud providers won’t build it, conflict of interest. Model providers won’t build it, same conflict. So we will.

For the builders AI pushed out

If you’ve watched good engineers get cut because a model could do the work, this one’s for you. Maybe it was your team. Maybe it was you. That happened, and pretending it didn’t is an insult.
Here’s the honest pitch. The same wave that displaced people created a control problem nobody has solved, and it needs exactly the kind of deep engineering the industry decided it could automate away. Kernel work. Hardware failsafes. Adversarial systems that hold up when something smarter than the last thing is trying to break them. It rewards the people who know how the machine actually works underneath the abstractions. You’ll build the thing that keeps humans in charge of the technology that took the job. That’s not consolation. That’s the point.

What you’ll own

The whole technical surface. Software control plane, edge hardware failsafe, and hardened control infrastructure that runs in a separate administrative domain from the systems it governs. CTRL sits outside the execution environment, at the boundary between the operational stack and everything an autonomous system must reach to act: identity, accounts, network, APIs, secrets, runtime, compute, power, hardware pathways, and actuators.
Three product wedges:

• Digital control plane. Independent control over digital autonomous systems, financial services first.
• Hardware failsafe. A fieldable module that mirrors the control layer into physical systems, down to power and actuation.
• Hardened control infrastructure. Dedicated hardened environments, sovereign over time.

The enforcement primitives are simple by design: throttle, degrade, isolate, pause, stop, recover, override. What makes them useful is the behavioral ontology and cross-layer observability underneath, which decide when to fire which one and at which layer. Kernel, hypervisor, bare metal, across digital and cyber to physical and hardware autonomy. We kill processes, block syscalls, suspend workloads, and recover to a safe state below the application layer, even when the guest environment is degraded or compromised. There is real application and network work too, and it matters. But if you love writing production code close to the metal, that’s the center of gravity.

How we work

Everyone is an engineer and a builder first. No PMs translating requirements. No scrum masters. No architects handing designs down. Engineers own the translation from mission to production, and you’ll set that standard as the first technical leader. Workflows are AI-native by default, and every engineer ships at AI-accelerated velocity. You’re a working CTO, hands on keyboard, not a manager who used to code.

The hard part

We will be one of the most actively targeted systems in the era of autonomy, by humans and by AI. The threat model is adversarial from the first commit. If we’re the layer that takes the keys back, everything trying to keep them is coming for us. Zero trust from first principles, not as a slide.
The bar the work has to clear:

• Sub-100ms policy decision latency. A control decision that arrives late is not a control decision.
• Sub-200ms hardware failsafe latency for physical interventions.
• NIST 800-53 and IL5/IL6 alignment as design constraints, not afterthoughts.
• Speed and security are the product. Both are non-negotiable. ### What we’re looking for
• Track record over tenure. You’ve built and shipped systems that had to work, at kernel or hardware depth, under real production or mission pressure. We don’t care whether that took you five years or twenty. We care that you can do it again, at the head of an engineering org, in front of adversarial customers and adversarial systems.
• Prior founding or early-stage technical leadership. You’ve built an org, not just joined one.
• Real depth at kernel, hypervisor, and bare-metal layers, with range across digital to physical.
• Adversarial security instinct. You assume something is trying to break it, because something is.
• Executive-grade communication. You’ll represent the technology to investors, defense customers, and Congress, and sound credible in every room.
• Builder of teams and standards. You set the culture and hire the bench.
• Willing to travel 10 to 25%.

Hard requirements: U.S. Person status and the ability to obtain a Top Secret clearance.

Why this seat matters

The tailwinds are real, not aspirational. On the defense side, we have active pathways across multiple Department of War program offices and research organizations, top cover from former senior military leadership, and congressional visibility at the committee level.
On the commercial side, the pull is just as real. A regulated financial services design partner is budgeting seven figures in 2026 specifically for the control layer. A top-tier global systems integrator is positioned to carry the product into regulated enterprise. Multiple large financial and infrastructure operators are in active conversation about the control problem in production AI. More lining up behind them as the demo matures.
The market is moving with us: a $13.4B DoW AI budget and $49B in defense tech venture funding in 2025, and every regulated enterprise now living with AI in production and no answer for the day it acts on its own.
The CEO is Army Special Operations and the Intelligence Community, with mission-critical systems experience at Palantir and Axon. His prior venture, Native, scaled to $2.5M ARR and 35 people, won five U.S. Air Force contracts, and did it on a fraction of the burn its peers were running. He’s hiring a CTO so he can stay on company-building while you run the technical org.
This is a true co-founder seat with meaningful equity. If you want to own the technology that keeps humans in charge of autonomy, this is the one.

CTRL is an equal opportunity employer. We consider all qualified applicants without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, disability, veteran status, or any other protected characteristic. This role requires U.S. Person status and the ability to obtain and maintain a U.S. Top Secret security clearance.