Applied Intuition's WarshipOS: 200 Roles vs a Pool of 3,083

Applied Intuition (not Applied Compute, the StartupHub.ai post conflated the brands) just formalized WarshipOS as a full product line and opened a Head of Engineering req in Ann Arbor at $250K to $350K base. Around it sits roughly 200 adjacent reqs across Sunnyvale, DC, Stuttgart, Munich, and Bangalore. The problem is not the offer band. The specialist pool it needs is already being drained by Anduril, Shield AI, Saronic, and Palantir at the same time.

The correction first: this is Applied Intuition, not Applied Compute

Applied Intuition, the $15B autonomy company founded in 2017, owns the DefenseOS and WarshipOS product line. Applied Compute is Rhythm Garg's ex-OpenAI post-training startup, an entirely different company. The Ashby listing is unambiguous: "We are hiring a Head of Engineering to build and lead the DefenseOS engineering organization, with an initial focus on building WarshipOS."

That matters for sourcing because the two companies pull from opposite pools. Applied Compute pulls from LLM post-training and RLHF. Applied Intuition pulls from automotive AV, robotics, and embedded systems, the same pool feeding every serious defense-tech company in 2026. If you are running defense tech recruiting against the wrong company brand, your outbound list is wrong by 100%.

What WarshipOS actually needs, in one screen

WarshipOS is a zonal compute and autonomy stack for surface vessels, built around four engineering pillars borrowed almost verbatim from automotive E/E architecture. The JD is explicit about the quadrant, and it is the exact skill mix the rest of the sector is fighting over.

The four pillars, from the listing:

1. Zonal compute architecture, with "location-based zone controllers and centralized compute that replace federated, domain-based designs"
2. Data transport, API, and SDK layer connecting sensors, compute, and downstream consumers
3. AI/ML perception and sensor fusion platform
4. Modern SIL/HIL test with a digital twin

Read that list twice. "Zonal architecture" and "zone controllers replacing federated domain-based designs" is not defense language. It is Bosch, ZF, and Continental language, ported to a hull. Applied Intuition's Ann Arbor, Stuttgart, and Munich offices are not a coincidence. They are a deliberate arbitrage on the automotive AV talent glut left over from the Zoox, Cruise, and Waymo contractions.

The pool is 3,083. Anduril alone has 2,809 open reqs

In Refolk's index of professional profiles, the US pool of Senior, Staff, and Principal software engineers with both computer vision and sensor fusion in their history is 3,083 people. Anduril has 2,809 active reqs in 2026 according to Revelio Labs, a 92.9% jump from 2025. That single company's demand equals 91% of the entire specialist pool before anyone else shows up.

91%

Now add the rest of the demand side. Applied Intuition's ~200 WarshipOS and adjacent reqs. Shield AI's autonomy roles in San Diego. Saronic's post-$600M raise hiring for uncrewed maritime, which overlaps WarshipOS's exact scope. Palantir's forward-deployed engineering funnel. Zoox and Waymo trying to poach their own alumni back. The specialist pool is not a rounding error against demand. It is smaller than the reqs.

The math does not close with a job board. It closes with poaching, and poaching only works if you can name the 3,083 people before your competitors do.

The clearance moat is smaller than the headlines suggest

Day-one active clearance is not the actual bottleneck for most of these reqs. It is a filter recruiters over-apply, and it is why the "cleared engineer sourcing" search bar is the wrong entry point for dual-use defense tech.

Here is the split that actually matters:

• Day-one cleared required: concentrated at the primes (Lockheed, RTX, Northrop) and specific IC-facing roles inside Palantir
• Sponsor post-hire: Anduril, Shield AI, Saronic, Skydio, and Applied Intuition's DefenseOS for most engineering roles
• Never cleared: the majority of dual-use platform engineering, including a large share of WarshipOS's zonal compute and perception work

DCSA's 90th-percentile processing time for Tier 5 (Top Secret) is around nine months, with roughly 19,000 Tier 5 cases in the pending queue as of mid-2025. Tier 3 (Secret) runs 60 to 90 days. Those are real numbers, but they apply after you already made an offer. They do not gate whether someone can start writing perception code on WarshipOS.

The practical reframe: sourcing strategy should target AV and robotics alumni pools first, then filter for clearance-eligible (US citizen, no disqualifying history) rather than actively cleared. That single change roughly doubles the addressable pool for most Applied Intuition reqs.

Day-one cleared is a filter recruiters over-apply. Clearance-eligible plus CV plus sensor fusion is the real query.

Why senior candidates clear slower than mid-level

The SF-86 itself is not usually what stalls a senior engineer. The FBI human-reference and field-interview portion is. The bureau's name-check backlog dropped from 42,000 to 20,000 after a new prioritization tool, but the reference work stayed manual.

Fifteen-plus-year work histories mean more former managers to interview, more foreign contacts to verify, more addresses to check. A staff engineer with a decade of history clears materially faster than a principal with twice that. If you are staffing WarshipOS on a 12-month schedule, your cleared-role hiring plan should skew slightly younger and lean on post-hire sponsorship for the seniors you actually need for architecture decisions.

Where the 3,083 actually work today

The named entities inside the CV plus sensor fusion pool are not evenly distributed, and that concentration is what makes targeted sourcing tractable. Refolk's index surfaces the top employers of the specialist pool: Zoox, Google (Waymo), and Waymo itself, followed by Cruise alumni now scattered across Applied Intuition, Nuro, and back into big-tech AI teams.

That is why Applied Intuition's geographic footprint reads like an AV alumni map:

• Sunnyvale HQ: Zoox, Waymo, Nuro proximity
• Ann Arbor: May Mobility, Ford AV, Argo alumni still in the region
• Stuttgart and Munich: Bosch, Mercedes, BMW, ZF zonal architecture teams
• San Diego: overlap with Shield AI and Qualcomm autonomous systems
• DC and Ft. Walton Beach: primes and SOCOM adjacency for cleared work
• Bangalore: offshore leverage for the perception and SDK layer

Naming a company is easy. Naming the specific 40 to 80 people inside that company who match a WarshipOS zonal-compute req is the actual job, and it is the exact gap Refolk closes. You describe the person in plain English (staff engineer, C++ or Rust, zonal or centralized compute experience, ex-AV, US-based) and get a ranked shortlist across GitHub, LinkedIn, and the open web.

The four sourcing plays that actually move the needle

Given the pool math, the productive plays for Applied Intuition hiring (or anyone competing with Anduril's talent pool) collapse to four. Everything else is theater.

1. AV alumni harvest, not cleared job boards. Query the CV plus sensor fusion senior pool by past employer (Zoox, Waymo, Cruise, Argo, Nuro). Filter for US-citizen and clearance-eligible. Ignore whether they are actively cleared. This is where most of the reachable pool lives.
2. SIL/HIL and factory-integration specialists. The sector's constraint is now execution, not invention. Recruiters who can source SIL/HIL test engineers and manufacturing integration leads will out-perform SWE-only recruiters. Smaller pool, less contested.
3. Zonal-architecture Europeans on L1 or O1 paths. Stuttgart and Munich zonal-controller engineers at Bosch, Continental, ZF, and Vector are the direct source of the WarshipOS architecture pattern. Applied Intuition's German offices exist to intercept them locally, but US relocation for the DC and Ann Arbor teams is realistic on L1B transfers.
4. Golden Dome adjacency mapping. Anduril's Golden Dome consortium partners (Impulse Space, Inversion Space, K2 Space, Sandia National Laboratories, Voyager Technologies) are second-order poaching targets. Their embedded and perception engineers already accept defense-tech work culturally, which cuts the "will they take a defense job" objection.

The $14.2B backdrop makes this permanent, not cyclical

US defense-tech equity funding hit $14.2B, compared to $2.48B in Europe, and category-wide venture into military, national security, and law enforcement passed $14.6B, breaking the prior $9.6B annual record. Anduril took $2.5B of that in June. Saronic took $600M. Applied Intuition sits on a $15B valuation, and Anduril's contract book now includes the $22B IVAS program takeover, a $100M NGC2 award, and a $23.9M Bolt-M contract for the US Marine Corps.

Capital is not the constraint. As one sector read puts it, "execution, not invention, will determine returns." Execution means engineers. Engineers means the 3,083. The pool does not grow on a capital-cycle timeline. It grows on a career-cycle timeline of five to seven years. Every defense-tech company hiring against WarshipOS in 2026 is fishing the same pond, and the pond is not refilling before the reqs close.

FAQ

Is Applied Compute the same as Applied Intuition?

No. Applied Intuition is the Sunnyvale autonomy company, founded 2017, valued at $15B, that owns DefenseOS and WarshipOS. Applied Compute is a 2025 AI post-training startup founded by ex-OpenAI's Rhythm Garg. The StartupHub.ai jobs page conflated the two brand names. The Ashby listing for the WarshipOS Head of Engineering role is at Applied Intuition, and Applied Intuition's ~200 reqs are what compete with Anduril's talent pool.

How many US engineers can actually build the WarshipOS stack?

In Refolk's index, the US pool of Senior, Staff, and Principal software engineers with both computer vision and sensor fusion is 3,083 people, roughly 5.1% of the 60,897 embedded plus C++ senior pool. Filter further for zonal-compute or SIL/HIL experience and the addressable number for any single Applied Intuition req is in the low hundreds. That is why targeted sourcing beats broad outreach in this sector.

Do WarshipOS engineers need active clearance to start?

Mostly no. Day-one active clearance is concentrated at the primes and specific IC-facing roles. Applied Intuition, Anduril, Shield AI, Saronic, and Skydio typically sponsor clearances post-hire. The real bottleneck is being clearance-eligible (US citizen, no disqualifying history) plus having the CV, sensor fusion, and embedded systems background. Recruiters who filter for actively cleared cut their addressable pool by more than half unnecessarily.

Where should Applied Intuition source from first?

The Zoox, Waymo, Cruise, Argo, and Nuro alumni networks, followed by Bosch, Continental, and ZF zonal-architecture teams in Stuttgart and Munich. The WarshipOS JD language is automotive E/E architecture ported to a hull, so the natural talent overlap is AV, not defense. Applied Intuition's Ann Arbor, Stuttgart, and Munich offices are deliberately positioned to intercept that pool locally.