Neko Health's $700M NYC Landing: The US Imaging-AI Pool Is 74

On July 15, 2026, Neko Health closed a $700M Series C led by Lightspeed at a valuation near $7B and confirmed a Manhattan clinic opens later this year. The headline is the money. The problem for anyone hiring against it is that the US engineering pool Neko wants, imaging hardware plus applied ML on the same resume, is roughly 74 people. Prenuvo and Forward were already draining it.

Neko's US landing, in one paragraph

Neko Health raised $700M in Series C funding on July 15, 2026, led by Lightspeed Venture Partners and co-led by O.G. Venture Partners, to open its first US clinic in New York City. The round included Atomico, General Catalyst, Lakestar, Liberty City Ventures, Positive Sum, and BDT & MSD, with angel checks from Mark Zuckerberg and Priscilla Chan, Tim Ferriss, Maria Sharapova, will.i.am, and OpenAI. Since launching in 2023, 350,000 people have registered for a Neko scan and 100,000 have received one, at £299 in the UK and 2,750 SEK in Sweden. US pricing has not been announced, but a Manhattan clinic sets a wage floor the scan price has to clear.

The company was co-founded by Hjalmar Nilsonne and Spotify's Daniel Ek. It has clinics in London, Birmingham, Manchester, and Stockholm. The Series C follows a January 2025 Series B at $1.7B, so the valuation moved roughly 4x in eighteen months. That is the fundraising story. The hiring story is smaller and stranger.

The pool is 74, not 200

In Refolk's index of professional profiles, exactly 74 US-based people list both medical imaging and machine learning as skills. That is the entire supply-side universe Neko Health hiring, Prenuvo, Forward, Ezra, and Midjourney's rumored body-scanner effort are competing over for individual contributor engineering work. Broadening the query to "medical imaging AI" and restricting to Senior, Director, or VP seniority produces 94 profiles. Leadership is thinner in absolute terms than IC talent, which is the inverse of most engineering markets.

74

The "200-ish" figure that gets thrown around in preventive-health recruiting decks is the addressable universe if you count radiology PhDs who have touched a CNN, or ML engineers who once shipped a DICOM parser. The intersection that Neko actually needs, people who can debug an imaging sensor stack and train a segmentation model in the same week, is smaller by half.

Why the intersection is so narrow

Medical imaging engineering is a career track that starts in physics, EE, or biomedical engineering and lives inside device companies with long product cycles. Applied ML is a career track that starts in CS and lives inside consumer or infra teams with short product cycles. The two cultures barely overlap, and the credential ladders point in opposite directions. Anyone who has bridged both is either a mid-career device engineer who taught themselves PyTorch, or an ML engineer who spent three years at a Philips or a Neuralink.

NYC gives Neko almost no local sourcing edge

Neko's Manhattan clinic is a demand-side bet on wealthy self-pay patients, not a supply-side one. Roughly 2 of the 74 core-pool profiles sit in NYC metro. Boston has about 2. The Bay has about 2. The profile is not concentrated in any tech hub; it is scattered across cities where big device companies happen to have offices.

The practical read for NYC health tech sourcing: Neko's recruiters will spend most of their time relocating candidates or building a remote-first ML team while the clinical staff, scan operators, and consumer-facing engineers cluster in Manhattan. Anyone selling a "we have to be in New York" narrative to imaging-AI candidates is going to lose to a Vancouver competitor that ships remote offers on day one.

Prenuvo's Vancouver HQ is Neko's biggest recruiting gift

Prenuvo's senior AI and staff engineering reqs are concentrated in Vancouver and Toronto, with only a handful of US postings in Seattle and South Carolina. That is the single largest structural advantage Neko has in the US preventive-health startup recruiting fight. Any Prenuvo ML engineer who wants US work authorization without the TN-visa friction of a Canadian employer is a warm lead the moment the Manhattan clinic posts a job description.

Prenuvo's public ML infrastructure listings describe designing inference infrastructure and migrating from AWS SageMaker to AnyScale and Ray for distributed execution. That is a specific signal recruiters should screen on directly. If you are sourcing Prenuvo competitors talent, "Ray + SageMaker migration + medical imaging" is a five-word Boolean that returns a very short list, and most of that list has a Prenuvo badge on it.

Neko isn't fighting Prenuvo for talent. It's fighting Philips, Bayer, and Neuralink, and pretending it's fighting Prenuvo.

The real competitors are device incumbents, not startups

The employers currently holding this profile in Refolk's index are, in rough order: Philips, Bayer, Lantheus, Merge Healthcare, Neuralink, Noah Medical, Riverain Technologies, and the Department of Veterans Affairs. Preventive-health startups collectively account for a rounding error on the employer list. If you build your Neko Health hiring sequence around poaching from Forward and Prenuvo, you will run out of candidates in a week.

The correct poaching map has three tiers:

1. Medical device incumbents with active ML groups. Philips, Bayer, Lantheus, Merge Healthcare, and Siemens Healthineers. Long tenure, mature imaging chops, ML capability arriving in the last 3 to 5 years.
2. Big Tech health arms. Microsoft Health and Life Sciences, Google Health, and Neuralink for the hardware-plus-ML variant specifically.
3. Public-sector and research feeders. The VA imaging AI teams, the Pediatric Accelerated Intelligence (PAI) Lab at Children's National Hospital, UTHSCSA, and MICCAI, RSNA, and SIIM conference alumni.

The third tier is where Neko's competitors are not yet looking. VA imaging researchers in particular are underpriced because government pay bands compress senior salaries, and the transition from federal service to a Series C startup is a well-worn path with predictable comp deltas.

This is the specific friction Refolk is built for. You describe the person in plain English, including the awkward "and" that Boolean search cannot express (imaging hardware AND applied ML AND US-based AND not already at Prenuvo), and get a ranked shortlist across GitHub, LinkedIn, and the open web. In a 74-person pool, the recruiter who reaches the right eight people first wins the quarter.

Vertical integration multiplies the hiring problem

Neko builds its imaging hardware, clinical software, and clinic experience in-house, which means it needs a rarer variant of the profile than radiology-model shops do. Vertical integration is the foundation of Neko's product story, and it is also the reason the hiring math is harder than a $700M round makes it look. A radiology AI startup can hire from the whole 74-person pool. Neko can only hire from the subset that has touched imaging sensor hardware, optics, or signal processing, not just DICOM ingestion.

That subset is closer to 20 than 74. The employers in the poaching map above who actually train this variant of engineer are Philips, Bayer, Neuralink, and a handful of academic labs. Everyone else on the list has ML engineers who consume imaging data but do not design the sensor stack that produces it.

What to screen for on a resume

• Experience with specific imaging modalities: MRI pulse sequences, CT reconstruction, ultrasound beamforming, or optical coherence tomography.
• Hands-on FPGA or DSP work alongside PyTorch or JAX.
• Publications at MICCAI, IEEE TMI, or SPIE Medical Imaging, not just NeurIPS.
• SIIM hackathon participation, which is a stronger signal than most ML conference attendance because the cohort is small and the projects are constrained by real clinical data.

The Prenuvo, Forward, Ezra, Neko comparison recruiters actually need

The four preventive-health startups fighting for the same imaging-AI engineers do not compete on the same axes. A sourcing pitch that works on a Prenuvo engineer will not work on a Forward engineer, because their product philosophies attracted different people.

Ezra is the interesting entry on that list for anyone doing NYC health tech sourcing. It already has NYC gravity and a similar scan-to-read workflow, and its engineers have been quietly available since the last funding cycle tightened. Neko's Manhattan opening will accelerate that motion.

The 90-day sourcing playbook

Given a 74-person core pool and three well-funded competitors, the sourcing motion for medical imaging AI engineers has to look different from a generic ML search. Here is what actually works in a pool this small:

1. Map the whole pool before you message anyone. In a 74-person universe, one badly targeted InMail burns 1.4% of your addressable market. Tools like Refolk let you scope the entire pool in a single query before you write a single message.
2. Segment by hardware exposure, not seniority. The hardware-plus-ML subset is the constraint. Sort candidates by whether their last two employers built sensors, not by title.
3. Lead with the vertical integration story, not the valuation. Engineers who left Philips for a startup did not leave because Philips was small. They left because they wanted to ship. Neko's in-house hardware plus software plus clinic footprint is a stronger pitch than "$700M round."
4. Warm the Prenuvo Vancouver cohort with US work-authorization messaging. Do not pitch culture. Pitch geography and green cards.
5. Watch MICCAI 2026 attendance lists and SIIM hackathon cohorts. These are where the 74 people gather, and both events happen inside the next Neko hiring cycle.

Refolk's index is where the first two steps live. Ask for "US-based engineers with both medical imaging hardware and deep learning experience, not currently at Prenuvo or Forward" and you get a shortlist that fits on one screen, which is exactly the right size when the pool is 74.

FAQ

How many US engineers can Neko Health actually hire from?

The core pool of US-based engineers with both medical imaging and machine learning skills is 74 profiles in Refolk's index. Broadened to "medical imaging AI" and filtered to Senior, Director, or VP seniority, the pool grows to 94. That is the total addressable universe for Neko Health hiring, Prenuvo, Forward, and Ezra combined at the IC and senior IC layer, before you subtract candidates already off the market or unwilling to move.

Does opening the clinic in NYC give Neko a sourcing advantage?

No. Roughly 2 of the 74 core-pool profiles sit in NYC metro, the same density as Boston or the Bay Area. The Manhattan clinic is a demand-side bet on wealthy self-pay patients, not a talent-supply move. Neko will need to relocate candidates or build a remote-first ML team while clinical and consumer-facing staff work onsite in New York.

Who is Neko Health actually competing with for engineers?

The direct startup competitors are Prenuvo, Forward, and Ezra, plus a rumored AI body-scanner effort from Midjourney. But the employers actually holding the profile Neko needs are medical device incumbents (Philips, Bayer, Lantheus, Merge Healthcare) and Big Tech health arms (Microsoft Health and Life Sciences, Google Health, Neuralink). Preventive-health startups are a rounding error on the employer list.

What is the fastest way to source Prenuvo competitors talent in this pool?

Skip title search and skill search entirely. Query the intersection directly: US-based, medical imaging plus ML, not currently at a preventive-health startup, with hardware exposure in the last two roles. Refolk handles that query in plain English, returning a ranked list from the 74-person pool with current employer and modality tags, which is faster than any Boolean string you can write against LinkedIn Recruiter.