Neko Health's $700M NYC Clinic Is Chasing a Hybrid Pool of 12

On July 15, 2026, Daniel Ek's Neko Health closed a $700M Series C at a ~$7B valuation to open its first U.S. clinic in New York. The press cycle framed this as a capital story. It is actually a sourcing story, and the arithmetic behind it is uncomfortable: the hybrid engineer Neko needs (medical imaging ML plus FDA-experienced hardware plus willing to sit near a NYC clinic) exists in the low double digits nationwide.

What Neko actually raised, and why the money isn't the constraint

Neko Health raised $700M led by new backers Liberty City Ventures, Positive Sum, and BDT & MSD, alongside existing investors Atomico, General Catalyst and Lakestar. The round follows a $260M Series B in January 2025 and values the company at nearly $7B. The angel list reads like a Vanity Fair party: Mark Zuckerberg and Priscilla Chan, Maria Sharapova, will.i.am, Thierry Henry, Jimmy Iovine, Tim Ferriss.

The traction is real. More than 100,000 people have taken a Neko scan, 350,000 more are registered or waiting, and three in four returning members whose scans flagged a serious condition now have it under control. The team is 1,000+ across Europe and the U.S., with hubs in Stockholm, New York, London, and Berlin. LinkedIn currently lists 94 open Neko roles worldwide, and the pipeline runs on Ashby.

None of that solves the actual bottleneck. Neko owns the full stack: its newest Stockholm clinic uses the next generation of its Derma-2, Echo-2, and Spectrum-2 medical devices, built in-house. Every senior hire needs to speak two languages at minimum. That is not a money problem. That is a supply problem.

The hybrid pool, in one table

In Refolk's index of U.S. professional profiles, the target hybrid intersection is measured in dozens, not thousands. Here is the honest picture.

Read that fifth row twice. Neko chose New York for the flagship clinic, but the U.S. medical-device hardware bench clusters on the West Coast and in Boston. Every senior hardware hire is either a relocation package or a bi-coastal engineering org from day one.

12

Why the intersection is the real bottleneck

Deep-learning engineers are abundant. Hardware engineers are abundant. Clinicians are abundant. The intersection is 350 nationwide, and Neko cannot substitute one leg for another because the product is a physical scanner running proprietary ML on regulated hardware.

That is the trap of owning the full stack:

• The ML side wants people who have shipped models against DICOM, ultrasound, or dermatology imagery. Most FAANG deep-learning engineers have not.
• The hardware side wants FDA 510(k) or PMA experience, ISO 13485, IEC 62304. Most consumer-hardware engineers at Apple or Meta have never touched a regulatory submission.
• The clinical side wants radiologists, cardiologists, and dermatologists who will operate inside a scan-first workflow, not a hospital.

Any given candidate typically has one of these. Neko needs teams where two of them overlap in the same person, or where three of them overlap in the same 10-person cell. That is the exact search you cannot express in Boolean, which is why plain-English search matters here: describe the hybrid you actually need and let the system rank, instead of AND-ing yourself down to zero. That is where Refolk earns its keep for healthtech recruiting NYC teams running against a 12-person local pool.

Who already owns the bench Neko wants to poach

The top U.S. employers of "Medical Imaging" + "Deep Learning" talent, in Refolk's index, are Amazon, Siemens Healthineers, GE HealthCare, Google, Meta, Apple, and Columbia Health Policy. That mix tells you exactly what Neko's offer sheet has to look like.

Two things fall out of this list:

1. FAANG holds roughly 7 of the top 10 employer slots for this cohort. Neko is not competing on healthtech comp; it is competing on Meta and Apple total comp. A $7B valuation gives them the equity story, but recruiters should model 2x market base as the floor for senior ML-imaging hires.
2. Siemens Healthineers and GE HealthCare are the two incumbent imaging giants with concentrated benches. They are also slow, structured, and have made a decade of hires that would love a startup exit. That is where the warm-intro sourcing actually pays off.

A $7B valuation gives Neko the equity story to outbid Meta. It does not create a single new hybrid engineer.

The takeaway for anyone running Neko Health hiring against this list: forget the "top of funnel" volume metaphor. There is no top of funnel. There are 350 people, and every one of them has a recruiter Slack thread about them already.

Ezra is the sleeper competitor, and its alumni map is the whole game

Ezra, founded in 2018 and based in New York, was acquired by Function Health in May 2025. That single fact reshapes Neko's NYC sourcing plan. The most obvious source of Neko-compatible talent (people who have already chosen to work on AI-driven preventive imaging in New York City) is now inside a well-capitalized competitor with an existing brand and a clinician network.

Any preventive health startup jobs strategy that doesn't map Ezra's current staff and alumni is malpractice. Specifically:

• Ezra's ML engineering staff (small, but hand-picked for exactly Neko's use case).
• Ezra's alumni from 2019-2024, many of whom moved to adjacent radiology-ML startups or to Siemens.
• Emi Gal's extended network, the Ezra founder, a software and ML engineer who has spent seven years recruiting for this exact intersection.

Then there is Prenuvo on the West Coast: 26 clinics across North America, Australia and Europe, with an affiliated radiology group employing over 100 radiologists. Prenuvo is ramping radiologist recruitment aggressively (they are recruiting at AOCR 2026 in Las Vegas in April). For clinicians, Prenuvo's playbook works. For engineers, it doesn't scale, because the underlying pool doesn't scale.

The non-obvious inversion: clinicians are the easy side

Founders tend to assume clinicians are the hard hire and engineers are the easy hire. In preventive imaging at scale, that is backwards.

Prenuvo's radiology group employs over 100 radiologists. Function Health, post-Ezra, is scaling similarly. The clinician pipeline for a scan-first preventive product is a solved problem: there are enough US-licensed radiologists, dermatologists, and cardiologists that a well-funded operator can staff a clinic network in quarters, not years.

The engineering pipeline is not solved. There are 350 people with the ML-imaging combination and 177 with the medical-device hardware combination. Very few overlap. Neko can staff a NYC clinic's clinical floor faster than it can staff the four-person ML team that keeps the Spectrum-2 pipeline shipping model updates. Founders and heads of talent should flip their assumption and budget accordingly.

What a realistic 12-month plan looks like

Here is a plan that respects the actual pool sizes rather than the raise size.

1. Map the 350 by hand. Not scrape. Map. Enrichment by employer, publications, patents, and last-move recency. This is where a query-first tool like Refolk collapses two weeks of Boolean into an afternoon: you ask for medical imaging ML engineers with FDA experience and get a ranked list across GitHub, LinkedIn, and the open web in one pass.
2. Segment by movability. FAANG-held candidates have 2x base floor and 4-year vesting. Siemens/GE candidates have real comp gaps and structural boredom. Start with the latter.
3. Accept bi-coastal engineering. The 177-person medical-device hardware pool is not moving to NYC for a Series C, no matter the equity. Open a Bay Area or Boston hardware hub in month two.
4. Own the Ezra alumni graph. Every hire, every ex-hire, every intern. This is a 50-person universe you should be able to name in a spreadsheet by end of week one.
5. Recruit clinicians in parallel, at volume. Copy the Prenuvo playbook: AOCR, ARRS, and RSNA presence, plus a radiologist referral bounty.
6. Track the Ashby board weekly. Neko's own job posts, and their competitors', are the highest-signal artifact in this market.

A hardware plus AI talent pool this shallow rewards the operator who treats sourcing as a research problem, not a funnel problem. The winners will spend the first 90 days building a named list, not running LinkedIn InMail campaigns against strangers.

FAQ

How many U.S. engineers actually fit Neko Health's core hybrid profile?

In Refolk's index, 350 U.S. professionals list both "Medical Imaging" and "Deep Learning" as skills, and 177 list hardware/firmware roles with "Medical Devices" experience. The intersection of those two groups (engineers who span imaging ML and regulated hardware) is a small subset of each. Realistically, Neko is fishing in a hybrid pool of dozens nationally, not thousands.

Why is NYC a hard location for this hire, given the raise size?

The U.S. medical-device hardware bench clusters in Mountain View, San Diego, and Boston, not New York. Only about 12 of the 350 U.S. medical imaging ML engineers sit in the NYC metro today. That is roughly 3.4% of the national pool, which means Neko either pays relocation on almost every senior hardware hire or runs a bi-coastal engineering org from day one.

Who is Neko really competing with for engineers?

Not other healthtech startups. The top employers of U.S. medical imaging ML talent include Amazon, Apple, Meta, Google, Siemens Healthineers, and GE HealthCare. Neko is competing with FAANG total comp on one side and incumbent imaging vendors on the other, plus Ezra (now inside Function Health) for the small subset of NYC-based, preventive-imaging-aligned candidates.

What's the fastest way to build a named list for a pool this small?

Skip Boolean. In a pool of 350, keyword search collapses under its own false negatives, because the actual signal lives in publications, GitHub commits, and job history phrasing, not skill tags. Query-first tools like Refolk let you describe the hybrid engineer you need in plain English and get a ranked shortlist across GitHub, LinkedIn, and the open web, so your first week is spent on outreach, not on de-duping spreadsheets.