[Interview] Ex-Samsung engineer couldn’t find a cure for his hair loss. So he built one with AI.

The morning Becon’s founder Park Min-suk realized none of the advice he’d been given was working, he did what many tech-minded Koreans do when confronted with a problem no product could solve: he built his own.

Park was in his late 20s when he noticed his hair thinning—gradually at first, then more visibly.

“I tried everything—oral medication, topical solutions, those French-import shampoos,” Park said in an interview with Korea Biomedical Review. “But I didn’t feel like anything actually understood my condition.”

Park, now 46, was working as an engineer at Samsung Electronics, developing image processing algorithms and working on system-on-chip designs for televisions. He wasn’t a dermatologist.

But in 2019, frustrated by how impersonal and ineffective most treatments felt, he pitched an idea through Samsung’s in-house accelerator: a machine learning system that could scan the scalp, analyze symptoms, and track whether any of it was working.

“I spent about a year and three months in incubation. We formed a five-person team inside the company and built the prototype,” he said.

The prototype worked. In 2020, Becon—short for "Be Confident"—spun out as a standalone company. 

Today, Park leads a 17-person team that builds precision diagnostics for the scalp. The company's IoT scanner and cloud-based AI platform are now used in more than 2,500 clinics, salons, and hospitals across 30 countries.

“Hair loss isn’t just cosmetic,” Park said. “It causes real stress. Once it starts progressing, it needs to be approached more medically.”

Rethinking what hair loss looks like

Becon’s device uses optical and UV imaging, temperature, moisture, and even odor sensors to build a data-rich picture of the scalp. One scan delivers real-time analysis of 12 skin and 12 scalp conditions—flagging follicle density, hair thickness, pore size, dead skin, sebum, and sensitivity in under three seconds. The results are then cross-checked against a database of 40,000 ingredients to guide the next steps, Park said, aiming to replace guesswork with data.

Becon’s system also cross-references a curated database of popular hair products, identifying which ingredients are likely to help—or harm—based on a user’s scalp profile. “We recommend accordingly,” Park said, “whether it’s oily, dry, seborrheic, or damaged hair that needs relief.”

Park says that the personalized approach sets it apart from the one-size-fits-all solutions that still dominate the industry. Minoxidil, for instance, helps by prolonging the growth phase of follicles. Finasteride works by reducing DHT, a hormone linked to hair thinning. Newer immunomodulators like ritlecitinib or baricitinib suppress inflammation in severe alopecia areata cases. But Park is careful not to overstate their universality.

“Some of those products include growth factors that promote hair growth -- it’s true,” he said. “But how long-lasting that growth is, or whether it’s actually sustained, varies a lot. It depends heavily on your scalp and skin type.”

He added, “Minoxidil isn’t exactly gentle. For people like me with sensitive or reactive skin, it really didn’t work well. It made my scalp itchy and slightly painful.”

Becon’s platform instead compares the healthiest and weakest parts of a person’s own scalp, identifying areas where hair density or thickness has dropped. That difference becomes the baseline for what the company calls the Hair Loss Index, a four-part score based on follicle density, hairs per follicle, strand thickness, and total volume. Over time, the index helps users determine if a shampoo, supplement, or treatment is worth sticking with.

In a clinical review of biweekly scans over 24 weeks, Becon said follicle density improved by 6.8 percent, volume by 6.9 percent, and hairs per follicle by 9.1 percent.

“It’s not just about visualizing dandruff or oiliness,” Park said. “It’s about creating a feedback loop. You try something, measure the response, then make better choices.”

The rise of scalp tech

Hair loss is a uniquely emotional condition—visible, stigmatized, and often misunderstood. In Korea, the market for hair loss care is projected to grow 8.4 percent annually through 2030. 

The country also leads the world in hair loss cosmetic patents, accounting for 42.9 percent of filings globally over the past two decades. Yet despite the surge of new products—from blood circulation sprays to scalp oxygenation serums—scientific validation remains thin.

Park is trying to change that. He speaks of “scalp zones” like a dermatologist, breaking down user dashboards that show regional health scores—forehead, crown, back, and sides—color-coded to reflect keratin buildup, odor, moisture, and follicle activity. Over time, these snapshots can distinguish product effectiveness from placebo.

“If your sebum level is 1.8 times the population average, we’ll flag it. If your sensitivity index is 28 percent and everyone else is at 8, that’s a red light,” Park said.

He also says Becon's algorithm now recognizes skin and scalp types across all ethnic backgrounds with 96 percent accuracy. With over two million data points logged as of early 2025, he said the startup is preparing to go deeper—offering population-level insights and working with biotech partners to develop new clinical-grade interventions.

A consumer version called WithBecon will launch this summer, allowing users to scan and monitor their scalp at home via a connected app. Park sees it as a way to bridge professional analysis and daily care. “It lets people manage their condition independently, but they can still stay connected to the professionals they’ve seen before,” he said.