How we built an AI unicorn in 6 years – TechCrunch

Today, Tractable is worth $ 1 billion. Our AI is used by millions of people around the world to recover from traffic accidents faster, and it also helps recycle as many cars as Tesla puts on the road.

And yet, six years ago, Tractable was just me and Raz (Razvan Ranca, CTO), two college graduates coding in a basement. Here’s how we did it and what we learned along the way.

Building on a new technological breakthrough

In 2013, I had the chance to get into artificial intelligence (more precisely deep learning) six months before it exploded internationally. It all started when I took a course on Coursera called “Machine Learning with Neural Networks” by Geoffrey Hinton. It was like being struck by love. Back then, to me, AI was science fiction, like “The Terminator”.

But an article in the tech press said the academic field was on the rise. With 100 times the size of training data sets and 100 times more computing power becoming available by reprogramming GPUs (graphics cards), a huge leap in predictive performance had been achieved in image classification a year earlier. This meant that computers began to be able to understand the contents of an image, as humans do.

The next step was to bring this technology into the real world. At the university, Imperial College London, together with much more qualified people, we have created a plant recognition app with deep learning. We accompanied our teacher into Hyde Park, watching him take pictures of flowers with the app and laughing with glee as the AI ​​recognized the correct plant species. This was previously impossible.

I started to devote every free moment to classifying images with deep learning. Still, no one was talking about it in the news – not even Imperial’s computer vision lab was a part of it yet! I felt like I was in a revolutionary secret.

Looking back, focusing narrowly on a branch of applied science undergoing a revolutionary paradigm shift that had not yet reached the business world changed everything.

Look for complementary co-founders who will become your best friends

I had already been rejected by Entrepreneur First (EF), one of the best incubators in the world, for knowing nothing about the technology. Having changed that, I applied again.

The last interview was a hackathon, where I met Raz. He was researching machine learning in Cambridge, passed the EF tech test, and published articles on reconstructing shredded documents and poker bots that can detect bluffs. Its rudimentary web page read, “I am looking for data-driven solutions to currently intractable problems. Now that sounded (and where we would get the name Tractable).

This hackathon we coded all night. The next morning he and I knew something special was going on between us. We moved in together and spent years side by side, 24/7, from waking up in Pantera in the morning to coding marathons at night.

But we wouldn’t have gotten to where we are without Adrien (Cohen, president), who joined us as our third co-founder just after our round of selection. Adrien previously co-founded Lazada, an online supermarket in Southeast Asia like Amazon and Alibaba, which sold to Alibaba for $ 1.5 billion. Adrien would teach us how to build a business, inspire confidence and hire world-class talent.

Find potential customers early so you can adapt to the market

Tractable started at EF with a head start – a paying customer. Our first use case was… plastic pipe welds.

It was as glamorous as it sounds. The pipes that carry water and natural gas to your home are made of plastic. They are connected by solders (melt the two plastic ends, connect them, let them cool and solidify again as one). Image classification AI could visually check people’s welding setups to ensure good quality. Above all, it was real value for a revolutionary AI.

And yet, in the end, they – our only paying client – stopped working with us, just as we were lifting our first round of funding. It was hard. Fortunately, the number of pipe weld inspections was too small a market to interest investors. So we explored other use cases: utilities, geology, dermatology, and medical imaging.