Jasmin Hume Is Teaching AI to Read Nature's Ingredient List

There is a charmingly old-fashioned problem hiding inside modern food science: nature has produced millions of proteins, researchers have recorded their sequences, and for most of them we still have only a foggy idea of what they can do. Somewhere in that molecular crowd may be a better emulsifier, a useful texturizer or the answer to a manufacturing headache. The difficulty is finding it before the product team retires.

Jasmin Hume has built her career around that gap between possibility and usefulness. At Shiru, the company she founded in 2019, machine learning narrows a vast search space, laboratory work tests the predictions, and production partners help turn the survivors into ingredients. It is less a magic wand than a very disciplined relay race. Code passes the baton to biology; biology passes it to manufacturing.

Hume's own path has the same cross-disciplinary rhythm. She is a protein engineer who became a food chemist, a founder who spent time inside venture capital, and an investor who still runs a laboratory-minded company. Her most useful talent may be translation: between molecule and market, scientist and customer, the idea on a screen and the material in a bag.

A Scientist Who Kept Leaving the Lab

Long before Shiru, Hume was already bothered by the walls around science. In 2012, as a doctoral candidate at what is now NYU Tandon, she spoke to middle-school students aboard the USS Intrepid. She put up a photograph of herself as a McGill University cheerleader and told the room she hoped it would break their stereotypes about engineers, scientists and cheerleaders.

It was a neat piece of stagecraft, but it also revealed a durable instinct. Hume wanted science to be visible, surprising and available to people who did not already picture themselves inside it. While pursuing her doctorate, she taught science and mathematics at a Bedford-Stuyvesant middle school as a National Science Foundation graduate fellow. She coached an after-school robotics team. She tutored high-school students. She later judged a youth invention contest alongside will.i.am and inventor Dean Kamen.

“Scientists won't come out of wherever they're working enough.”Jasmin Hume, speaking about public engagement in 2012

Her geography was as varied as her interests. Born in Gothenburg, Hume studied engineering at McGill in Montreal, pursued graduate study at Chalmers in Sweden, worked in a biomaterials laboratory for STMicroelectronics near Milan, and completed a doctorate in New York. At NYU, her research dealt with engineered proteins and biomaterials, including proteins designed to assemble into fibers. The raw material was microscopic; the ambitions were architectural.

Even then, entrepreneurship kept intruding. In graduate school she started BenchPals, an app concept intended to help researchers digitize laboratory and data management. She also spent a summer at Lux Capital, seeing technical companies from the investor's side of the table. Neither experience appears as a detour now. Shiru would eventually demand fluency in software, science and capital at once.

The Ingredient Is the Infrastructure

After NYU, Hume joined Hampton Creek, later renamed JUST, and rose to Director of Food Chemistry. The work made an industry bottleneck impossible to ignore. Consumer brands could dream up new products quickly, but they were often assembling them from a limited pantry of familiar functional ingredients. The glamorous product at the end of the aisle depended on obscure proteins upstream doing unglamorous jobs extremely well.

Hume saw white space. Rather than build another finished food brand, she would build ingredients for many brands. Shiru began as the enabling layer, a business aimed at discovering functional proteins from nature and getting them into forms that manufacturers could evaluate and use.

The distinction matters. AI can rank candidates, but a ranked list is not an ingredient. A protein still has to express efficiently, survive processing, perform its intended function, fit an economic production method and behave consistently in the customer's formulation. Biology has a dry sense of humor about beautiful slide decks.

This is why Hume talks about discovery and commercialization in the same breath. The algorithm is only one worker on the line. Shiru's platform, Flourish, links amino-acid sequences with predicted functional attributes. Its public-facing marketplace, ProteinDiscovery.ai, lets customers search, request samples and begin discovery without building an internal biotech operation from scratch.

When TIME named ProteinDiscovery.ai one of its Best Inventions of 2024, the recognition landed on that access model. The interesting move was not merely cataloguing proteins. It was giving food scientists a way to order samples and conduct their own exploration. Hume has compared the idea to an online marketplace for proteins, which is both an audacious metaphor and a wonderfully practical one. Search, click, test - with considerably more pipetting.

From Screen to Sack

The sequence of recent milestones shows a company moving from platform promise toward physical goods. Shiru launched ProteinDiscovery.ai in May 2024. That autumn came TIME's recognition and a $16 million Series B led by S2G Ventures. In April 2025, Hume announced that commercial production had begun for OleoPro and uPro, Shiru's first two AI-discovered ingredients.

OleoPro is a structured-fat alternative intended to reproduce functions manufacturers often seek from animal fats. uPro is a texturizing protein. Their significance is less poetic than useful: each exists to perform a job inside another company's product. They are proof that the discovery engine can leave the screen, survive scale-up and arrive where industrial buyers need it.

There is an operating lesson here for anyone building deep tech. Hume did not treat manufacturing as the dull chapter after discovery. Production constraints belong near the beginning, because a marvelous molecule that cannot be made economically is a research result, not a product. Shiru evaluates different production routes, including microbial fermentation and plant-based molecular farming, according to the protein and the application.

The same pragmatism appears in Hume's public advice. In a 2022 conversation about solo founding, she described the balance between an ambitious, visionary picture and the tools available today. That tension is the founder's daily weather: keep the horizon large, keep today's experiment specific.

“The world needs your impatience.”Jasmin Hume, reflecting on founders who challenge industry convention

A Platform With the Doors Open

By 2025, Hume's argument had widened. In an essay for the World Economic Forum, she wrote that democratized AI tools could accelerate discovery across ingredients and materials. The bottleneck, in her telling, was not simply better technology. It was access to that technology. She imagined an emerging brand in Chile using the same class of discovery tool as a multinational in Switzerland.

That aspiration connects the STEM mentor on the Intrepid to the CEO in Alameda. Both are asking who gets to participate. A searchable protein marketplace may be an enterprise product, but its underlying wager is that more capable people should be able to ask nature better questions without first assembling an enormous computational biology department.

Hume extends that connective role through XFactor Ventures, where she invests in women-led companies, particularly in climate and deep tech. In writing about the bioeconomy, she and fellow partner Michelle Egger emphasized infrastructure businesses: the less photogenic tools, systems and inputs that make an industry possible. It is the same taste visible in Shiru. Finished products get the shelf space. Enablers change what can reach the shelf.

Off duty, a public conference biography places Hume with her family and dog, exploring the Bay Area, traveling to see relatives in New York or Spain, or baking Swedish pastries heavily armed with cardamom. The detail is small but fitting. For someone preoccupied with functional ingredients, moderation in spice would be suspicious.

In 2026, Inc. recognized Hume on its Female Founders 500 list. Her response resisted the qualifier even while welcoming the company. She wanted the emphasis on founder, and on changing the industry's assumptions about how quickly discovery can move. It was consistent with the old cheerleading photograph: accept the room, question the box.

Shiru's next test is commercial repetition. One pair of ingredients demonstrates a path; a platform earns its name by walking that path again and again. Hume's advantage is that she seems constitutionally unable to leave the handoffs alone. She notices where knowledge gets trapped, where people stop talking and where a decade has been mistaken for a law of nature. Then she builds a door.