Lightly: Meet the Venture Leader Technology curating data to optimize machine learning

Lightly offers a solution focusing on data selection and active learning for computer vision. The startup's software specializes in accelerating machine learning model deployment time by reducing the volume and complexity of required training data. Lightly co-founder Matthias Heller will join nine other Venture Leaders on an investor roadshow in Silicon Valley in April.

Whom does your product or solution help, and how?
Our solution helps everyone using machine learning with visual data for example companies doing medical imaging, autonomous driving, or video analytics—especially for companies that are deploying their AI systems in production. Lightly is a great fit to make sure their systems always work.

How did you come up with the idea for your startup?
We saw in our own ML projects that curating high-quality data, the "food" of AI, was an incredibly time- and resource-intensive process. So why not make this efficient for ourselves and all other companies? We tested the market by conducting dozens of user interviews and sent our whitepaper to hundreds of potential prospects until we won our first customers.

What market are you addressing and what is the potential of your startup in that market?
We are focusing on the computer vision market because it's the largest ML market for unstructured data, accounting for over 30% of the total market size. At the same time, it's the most complex and expensive data to curate, so the value we provide is significant. The market for ML-based products is doubling everywhere and will soon hit trillions of dollars. That being said: the market for companies like Lightly enabling production scale AI will be in the billions at least.

What excites you most about your work right now?
Every morning I wake up I am excited to go to work. There's no other job I would want more – no other activity I would rather spend my time on.

The people are the key. My customers are some of the brightest researchers and engineers at companies like iRobot, Syngenta, and many more that are building cutting-edge AI products. At the same time, I'm lucky to have the best team I could wish for with the most talented but at the same time most fun people to work with.

How did you go about building your team?
We are looking at three things: talent, personality (culture fit), and ambition. At Lightly, being talented and ambitious is not enough. It's also about fitting in with the team and having the right values. We are also continuously looking at hiring people with complementary skills and perspectives. That's why we are continuously asking ourselves: what is missing in the current team? What can bring us to the next level? This is in my opinion a good approach for building a team that is capable of high performance but at the same time fun to work with.



What are your team's key achievements to date?
There are many different achievements but building a product that people appreciate and are willing to pay for is probably the most important one. Furthermore, getting accepted into Y-Combinator and being able to raise a seed round were key moments for our company.

How did you come up with the name of your startup?
That's actually a funny story. It was a very long and iterative process. Many people have forgotten that we started with the project name "WhatToLabel." It took us several weeks, if not months, to come up with that name. It was a very exhausting thing to do.

However, the final name makes a lot of sense. We are making the lives of many engineers easier and, therefore making them "lighter." Hence, Lightly.

Is there a key principle or value that guides you as you build your company?
Customer centricity. I see so many startups building great products but never talking to customers. Before we start building features, we make sure to talk to potential customers, if not sign contracts. In short: Don't build things that don't sell.

Who do you look to for guidance and mentorship?
I'm fortunate to have had great mentors in my professional and personal life. The biggest impact on my entrepreneurial journey was made by Dr. Felix Grisard, who, in my opinion, is one of the most impressive entrepreneurs in Switzerland. His view on how continuous change affects the economy all the time and gives founders opportunities to build companies had a profound effect on me. That's why I'm grateful that even today, I can reach out to him to discuss questions and challenges.

Tell us something people don't know about you.
That I spend my weekends with my dog, Joy, roaming the Swiss forests, hunting for truffles. Usually, the found "treasure" will then be cooked with fresh homemade pasta and enjoyed with a glass of red wine.

What is your favorite podcast and why?
Swisspreneurs. It's such an inspiring podcast with real-world, applicable advice. Silvan Krähenbühl and the whole team are doing an amazing job. I can recommend that every founder and future founder listen in.

ABOUT THE PROGRAM
The Venture Leaders Technology 2023 were chosen from over 150 applications by a jury of investors and technology experts. The roadshow in Silicon Valley, one of the world's most important tech hubs, provides a unique opportunity for entrepreneurs to accelerate their startups' expansion in the US while expanding their professional network and profiting from workshops with investors and industry leaders. This year's Venture Leaders Technology is co-organized by Venturelab and supported by Canton of Vaud, Canton of Zürich, EPFL, ETH Zürich, Kellerhals Carrard, and Rothschild & Co.