Noah Levy has spent a surprising amount of his career worrying about boxes. Not spacecraft, exactly, though his work has flown close to that word. The troublesome boxes sat on laboratory benches and inside test racks: expensive instruments, each with its own cables, drivers, interfaces and little domestic rituals. The radios under test might be headed for a rocket, capsule or satellite. The gear used to interrogate them could still behave like it had missed several generations of the internet.
This mismatch became Levy's chosen problem. In 2025, after years of radio-frequency engineering at SpaceX, he co-founded Nine Fives with Andrew Kurtz. Levy took the CEO role; Kurtz, a fellow aerospace hardware veteran, became COO. Their company makes RF test equipment for teams building radios into spacecraft, drones, phones and other connected machines. The pitch is plain enough to fit on a label: networked, driverless instruments with a browser interface and APIs that software can read.
Plainness is part of the point. Test equipment is infrastructure. It is meant to disappear into the productive rhythm of a lab, not become the lab's recurring drama. Levy had seen what happened when fast hardware programs met slow vendor workflows. Nine Fives describes him as the "squishy object" caught between them. It is an excellent phrase for a bad job: the human buffer absorbing incompatible speeds.
I am not a fan of modern RF test equipment from legacy vendors.Noah Levy, in his Y Combinator founder profile
An engineer shaped by signal paths
Levy's route into the test rack began well before rockets. At Cornell University, where he earned bachelor's and master's degrees in electrical and computer engineering, he worked on the university's autonomous underwater vehicle team. His acoustics project required the whole embedded stack: an algorithm for locating a sound source, a schematic, a printed circuit board and STM32F4 firmware for passive sonar.
The result offers a small, telling portrait of engineering taste. The board was the team's first successful iteration in six years. Its digital logic cost $12, down from more than $400. That number has none of the glamour of orbital hardware. It says something useful about Levy anyway. He was not merely making a system function. He was removing accumulated awkwardness until it became cheaper, clearer and repeatable.
His early internships moved across software and hardware. There was MakerBot, then GPU programming and hardware design at NASA's Jet Propulsion Laboratory, followed by hardware systems engineering at Intel. He also built an amateur-radio modem in a Cornell microcontroller course. The contraption simulated a modulator and demodulator in Matlab, then implemented them in C on a PIC32. Connected to a low-cost handheld radio, it managed roughly 440 baud. Humble throughput, handsome specificity.
SpaceX followed: first an RF engineering internship in 2016, then a sequence of full-time roles from RF engineer to senior RF engineer. Levy worked on the Falcon 9 radar altimeter and the Dragon 2 S-band transceiver. His Starship work included GPS and S-band radio hardware. Later, he worked on satellite programs and a phased-array calibration system. Nine Fives says he received six internal Kickass Awards during those years. The name sounds like a prop from the company launch video, but the repetition points to sustained responsibility.
The critical path has a COM port
A spacecraft radio is a triumph of constraints. It must survive vibration, temperature, radiation, power limits and long distances. The lab around it has a quieter set of constraints, and quiet constraints are easy to ignore. An instrument needs the correct driver. Documentation lives in a PDF, perhaps behind an account. Eight devices decide they all want the same numbered serial port. A useful test waits while an engineer negotiates with the tools intended to enable it.
Nine Fives turns those inconveniences into a product specification. Its instruments can connect over Power over Ethernet or USB. They advertise themselves on a network. A web interface runs in an ordinary browser. For automation, each unit exposes REST and SCPI endpoints. The former fits contemporary software habits; the latter keeps faith with an instrument-control standard engineers already use. The trick is not novelty for its own sake. It is compatibility without ceremony.
The first product family makes the philosophy tangible. A programmable attenuator changes signal strength. An SPDT switch routes a signal between paths. Each module can work alone. A data-and-power backplane assembles modules into a rack while reducing cable sprawl. NineVue, the company's web application, represents the rack configuration so that it can be maintained and versioned. A custom skill file gives coding agents structured context about the system.
This does not make RF testing easy. It removes chores that have little to do with radio physics. That distinction is the company's practical center. The measurement still matters. Calibration still matters. Signal integrity certainly matters. An engineer should be spending judgment on those things, not persuading a rack of instruments to introduce themselves to a computer.
A baseball bat meets product marketing
Hardware workflow pain is awkward to explain outside the lab. It lacks the immediate theater of a broken screen or a smoking motor. Levy and Kurtz solved the communication problem with slapstick. Their public launch film stages the frustration of old test equipment with props, formalwear, flying paper and one particularly assertive baseball bat.
During filming, Levy accidentally hit himself in the face with the bat. His LinkedIn post supplied the line that made the accident part of the product story: he and Kurtz were building Nine Fives because the mishap did not rank among his five painful RF test-equipment experiences. The joke works because it compresses years of niche aggravation into a picture anyone can understand. The lab may be specialized. Exasperation is universal.
The humor also reveals a founder comfortable being the demonstration dummy. Levy's public profile includes an undergraduate project titled "Being a gangster," described with one dry line: "Damn it feels good." Alongside courses in analog integrated circuits, electromagnetic fields, computer architecture and embedded systems, the joke lands as a useful pressure valve. His professional persona is technically dense without being solemn about itself.
Two people, one opinionated rack
Nine Fives entered Y Combinator's Spring 2026 batch as a two-person company. Small teams are common in software; they are more conspicuous when the catalog contains physical instruments, machined enclosures, RF paths and a modular chassis. The partnership divides the load along familiar aerospace lines. Levy brings RF product and test-system experience. Kurtz brings mechanical engineering, manufacturing and integration experience from SpaceX and Boeing.
Their shared advantage is not simply that both worked around spacecraft. They know the difference between a device that performs on a bench and a product that can be built, documented, connected and trusted repeatedly. Nine Fives backs its equipment with a 30-day money-back guarantee and a two-year warranty. Those promises are quieter than the launch film. For a new hardware vendor, they are part of the argument.
What another builder can steal
- Repeated irritation is research when you can name the failure mode precisely.
- Modernization can preserve a familiar standard while adding a cleaner interface.
- A physical module should make sense alone before a platform makes it more useful.
- Niche pain becomes memorable when the explanation is concrete and human.
Levy's transition from SpaceX to Nine Fives has the logic of moving one layer outward. First, build the radio hardware. Then, improve the instruments surrounding it. Next, make the rack legible to an engineer, a script and an AI coding agent. The ambition is embedded in the interfaces rather than inflated into a slogan.
There is also a pleasing loop back to the Cornell sonar board. In both stories, a complicated technical system carries a layer of inherited cost and friction. Levy's instinct is to inspect that layer until it yields. At Cornell, the evidence was a $12 piece of digital logic. At Nine Fives, it is an RF instrument that appears on the network, opens in a browser and explains itself to software.
A laboratory will never be free of surprises. Signals leak. Connectors loosen. Measurements disagree at inconvenient hours. Levy is betting that the instrument interface need not add suspense. For someone whose launch film involved a necktie, airborne paperwork and accidental contact with a baseball bat, the desired product experience is almost comically calm. Plug it in. Find it. Run the test. Let the drama belong to the video.