SoftBank has a special fund focused on AI that's separate from the Vision Fund, and it just backed its first startup outside of Japan

Engineer.AI

Engineer.AI cofounders Saurabh Dhoot and Sachin Dev Duggal

• Custom software creator Engineer.AI has raised $29.5 million from Lakestar, Jungle Ventures, and SoftBank's AI-focused Deepcore fund.
• Engineer.AI makes it easy for people with modest budgets and limited technical know-how to commission a custom app, website, or other software from scratch.
• The startup uses AI to price its services and assign projects to its network of more than 30,000 engineers.

Engineer.AI, a startup that lets anyone create their own custom software, has raised $29.5 million in its first round of funding, led by Lakestar Ventures, Jungle Ventures, and SoftBank's Deepcore fund.

The Series A raise is notable for its size and for SoftBank's involvement. The Japanese firm has made waves by ploughing money into startups through its $100 billion Vision Fund, but it also has the newly established $52 million Deepcore fund.

Complimentary Tech Event

Transform talent with learning that works

Capability development is critical for businesses who want to push the envelope of innovation.Discover how business leaders are strategizing around building talent capabilities and empowering employee transformation.Know More

According to Engineer.AI's British cofounder Sachin Dev Duggal, this is Deepcore's first non-Japanese investment. SoftBank did not immediately respond to a request for comment.

Engineer.AI's key product is its Builder tool, which is targeted at people who want to create an app, ecommerce site, marketplace, or website for their business but don't have the technical know-how or budget to commission developers and designers to build something bespoke. Duggal gives the example of someone running a French ski school with 200 instructors wanting a site and app that connects their different systems, but unable to afford the requisite $500,000 for custom software.

Engineer.AI

What is particularly clever about Engineer.AI is that it is efficient. It reuses different building blocks of code over and over again across its clients' different apps or services, meaning there's often no need to code major, expensive features from scratch, such as search. A one-off feature that would cost $30,000 to build from scratch can be offered at a tenth of the price to customers, because Engineer.AI can keep reusing the same building blocks of code. Duggal said the most reused building blocks are built by the top 1% of engineers on the company's network of developers.

"It's an assembly line for making software," he said. "It's reusable building blocks, like car manufacturing, and a lot of automation."

As users navigate through Builder and pick out different features for their app, site, or other software, the site uses a neural network to come up with a price. Depending on how fast customers need the project and where they want it built, the software can be cheap or expensive.

See also: The 25 most valuable US startups that failed this year

Another aspect of Engineer.AI's service is "BuilderCare", a kind of extended warranty for the software. It costs extra, a little like Apple's AppleCare for iPhones.

And who is actually doing all the building?

Duggal says Engineer.AI has 31,500 engineers on its network and almost 100 developer shops.

Asked how the concept is different to asking freelancers on Upwork to code something cheaply, he said: "We don't work with freelancers, we buy excess capacity from other developer shops. And there is a lot more control over the individual doing the work.

"Secondly, unlike something like Upwork, we don't rely on a user rating... the bigger score is our AI that grades the [engineer's] code. We pick the developer that will work on [the software] on your behalf. And we guarantee quality and price."

Engineer.AI plans to use the new funding to scale up. It currently has split headquarters in London and Los Angeles.