BREAKING lululab reads 7 skin metrics from a single photo in ~10 seconds ●
Spun off from Samsung C-Lab, 2017 ●
CES Innovation Award honoree four years running ●
5,000,000+ skin data points and counting ●
Series C of ~$20M closed in 2022 ●
Netmarble & Coway in the partner column ●
BREAKING lululab reads 7 skin metrics from a single photo in ~10 seconds ●
Spun off from Samsung C-Lab, 2017 ●
CES Innovation Award honoree four years running ●
5,000,000+ skin data points and counting ●
Series C of ~$20M closed in 2022 ●
Netmarble & Coway in the partner column ●
The State of Play
A skin lab that fits inside a phone, a kiosk, and a countertop
Walk up to a Lumini Kiosk in a department store. Look into what appears to be an ordinary mirror. Ten seconds later it hands back a number for your pores, another for your wrinkles, and five more for things you did not know had numbers. This is lululab's ordinary Tuesday.
lululab Inc. is a South Korean company that turned skin into data. Its Lumini platform - an app, a retail kiosk, a home device, and a developer SDK - uses computer vision and deep learning to read a face across seven categories: wrinkles, pores, acne, sebum, melasma, redness, and the oil-moisture balance. From a single photo. The output is not a vibe. It is a score, and then a recommendation.
Skincare has always been an industry of confident guesses. lululab's whole proposition is to replace the confidence with a measurement.- THE CENTRAL IDEA, IN ONE LINE
The company is small - around 33 people - but its ambitions are not. It keeps a foot in Cambridge, Massachusetts, while running research out of Seoul and Daegu, and it sells its technology to brands, retailers, and clinics across more than one continent.
The Problem They Saw
The beauty aisle is a confidence game
Here is the uncomfortable thing about the global skincare market: most of it runs on self-diagnosis. People squint at their own reflection, decide they have "combination skin," and buy accordingly. The serum that worked for a friend becomes the serum you try next. It is a multi-billion-dollar industry built, in large part, on a hunch.
Dermatologists can do better, of course. But a clinical skin analysis means an appointment, a wait, and a bill. For a routine purchase decision, almost nobody books one. So the gap sat there in plain sight: skin health was measurable in principle and unmeasured in practice.
The data existed on everyone's face. There was simply no instrument cheap and fast enough to read it.- THE GAP LULULAB AIMED AT
That gap is the tension running through everything lululab has built. Every product is a different answer to the same question: how do you put an objective skin reading in front of an ordinary person, at the exact moment they are about to make a decision?
The Founders' Bet
From genome sequencing to your bathroom mirror
lululab spun out of Samsung Electronics' C-Lab in 2017, the 22nd company to graduate from the electronics giant's internal venture program. Its founder and CEO, Yongjoon Choe, took an unusual route to skincare. He studied bio-engineering at Cornell and worked on DNA genome sequencing in a Harvard-affiliated medical setting before concluding that the tools used to objectively measure skin were, frankly, embarrassing.
The bet was straightforward and slightly contrarian: that a camera plus a well-trained neural network could approximate what a trained eye does, but faster, cheaper, and at scale. Skin, it turns out, is a remarkably cooperative subject for computer vision - consistent features, repeatable patterns, lots of labelled examples once you start collecting them.
A gaming company funding a skincare AI sounds like a punchline. Netmarble wrote the check anyway.- ON THE SERIES B, LED BY A KOREAN GAMING GIANT
Investors came around. Netmarble, one of Korea's leading game developers, led the Series B in 2020, joined by L&C Bio, Global Medical Research Center, and CTK Investment. A Series C of roughly $20 million followed in 2022. Total disclosed funding sits near $19 million. Not a fortune by Silicon Valley standards - but enough to keep collecting faces.
The Product
One engine, four front doors
lululab did not build a single gadget. It built a skin-analysis engine and then wrapped it in whatever shape the customer needs - retail, home, or code. The deep-learning model is the constant; the packaging changes.
Lumini Kiosk
A self-service smart mirror for stores and clinics. One photo, seven skin scores, payment and product recommendation built in.
Lumini Home
A compact tabletop device that tracks a whole family's skin daily and connects to other care devices.
Lumini SDK
The engine as a service - drop skin analysis into any iOS, Android, or JavaScript app and let other businesses build on it.
Lumini App
The consumer-facing scanner that turns a phone selfie into a personalized routine and product shortlist.
The clever move was not the mirror. It was deciding the mirror, the app, and the SDK could all share one brain.- ON THE PLATFORM STRATEGY
The Mission
Make skin health measurable for everyone
Strip away the kiosks and the funding rounds and the mission is almost stubbornly simple: take something people have always guessed at and turn it into something they can actually read. lululab wants the objective skin score to be as ordinary as checking the weather - a number you glance at before you decide what to do.
That ambition is bigger than cosmetics. Skin is the body's largest organ and one of its most visible health signals, which is why lululab's keywords drift from "beauty" into "digital health," "remote patient monitoring," and "AI diagnostics." The same camera that sells you a moisturizer could, in principle, flag something a dermatologist should see.
If beauty was the way in, healthcare is the room they actually want to stand in.- ON THE LONGER ARC
Why It Matters Tomorrow
The mirror is learning
Phone cameras are getting better and stranger. The Spectricity partnership points at a future where a handset sees light the human eye cannot, and where a skin reading no longer needs a dedicated device at all. If that lands, lululab's engine stops being a thing you walk up to and becomes a thing you already carry.
There are reasons for the skeptic to stay skeptical. Self-reported accuracy is a claim, not an audit. The competitive field - Perfect Corp, Revieve, Haut.AI and others - is crowded, and "AI skin analysis" is fast becoming a checkbox feature rather than a moat. lululab's edge is its head start and its pile of five million labelled faces, which is the kind of asset that compounds quietly.
Whoever owns the most faces, fairly read, owns the recommendation. That is the game lululab is playing.- THE BET, RESTATED
So return to that department-store mirror. A few years ago it was glass, and the verdict it gave you was your own. Now it counts your pores in ten seconds and tells you, with a number, what to do next. lululab did not invent the desire to look closer. It just built the instrument - and handed it to anyone with a camera.
