Backing MindsDB — Automated Machine Learning for the Software Developer.
OpenOcean is unique among venture capital firms, due to our open source heritage. We were founded by the team behind MySQL, the open source database that helped define generations of software engineers (including me!) and which was eventually bought for $1b by Sun Microsystems.
Moreover, MariaDB — another highly successful open source database — was founded by OpenOcean’s founding partners, Monty Widenius, Patrik Backman, and Ralf Wahlsten.
This experience is priceless when assessing investment opportunities in the open source software industry, and when we met Adam and Jorge of MindsDB we were immediately very interested.
Open source was once infamously lambasted by Microsoft’s Steve Ballmer as a “malignant cancer.” This is about as far away as possible from Microsoft’s position today. In fact, its open source software contributions are massive. Alongside other major software vendors, Microsoft dominates the top-ranking projects on GitHub.
Not only is open source expanding its community reach, but it is also proven to be highly desirable commercially. There has been a surge in commercial open source funding. IBM’s $34b acquisition of Red Hat was a particularly notable transaction.
Patrik got to know both Adam Carrigan (COO) and Jorge Torres (CEO) of MindsDB at Web Summit in Lisbon. He quickly realised that they had a common belief in the enormous power of data and an enthusiasm about how open source software can deliver technology to the masses. Here were two founders, friends for many years and with a track record of joint endeavours, building an exciting open source product to democratise machine learning (ML) amongst software developers. They had our attention.
In 2017, Adam and Jorge had recognised that many businesses faced limitations in extracting meaningful information from their data. This was mostly due to their severe under-use of the power of artificial intelligence. There is a booming market in ML, a desire to build and train ML models and get them into production, but not nearly enough experts to satisfy the demand. Today, there are about 2.7 million data scientists and analytics professionals worldwide, but the market demand is for twice as many.
So Jorge and Adam designed a platform that would enable anyone to leverage the power of ML; to ask predictive questions of their data and get accurate answers. They called this platform MindsDB.
It was a sophisticated product from a technological point of view, addressing the growing problem in data science — the shortage of data scientists — by empowering software developers.
In a typical enterprise there are three major stakeholders when deploying ML: the domain experts, the data scientists, and the software engineers. Adam and Jorge knew that software engineers are fundamentally quantitative and are the gatekeepers of production, but rarely have ML expertise. They are clearly the people to empower in order to harness ML, working in conjunction with the domain expert.
There are more than 20 million developers in the world and, with MindsDB, they can plug the skills gap.
Furthermore, by reducing the reliance on specialist data scientists, MindsDB creates a much tighter relationship between the domain expert and software developer, to increase productivity.
MindsDB has built its open source AutoML tool for the world’s software developers and has established a unique position for itself. There have been many such tools — some of them highly successful — serving data scientists and business analysts (e.g., DataIku, RapidMiner, DataRobot, H2O). MindsDB, however, has built a best-in-class offering for developers and is building an open source developer community around it.
MindsDB has avoided a common pitfall for many open source businesses: it gets the balance right between free and open functionality versus paid and proprietary licensed functionality.
The open core of MindsDB’s project is well constructed, providing a valuable open source project to engage and grow a community of developers. It empowers them to start using neural networks and provides useful explanations of the output.
The monetised elements, the proprietary licensed functions, are well constructed to provide businesses with features so that models can be deployed into production with confidence. In the era of DevSecOps, much is already being asked of the software developer. DevSecMLOps must have powerful, reliable tooling that is a pleasure to use, if it is to succeed.
These are just some of the many reasons why we are delighted to support this dynamic team on its journey. OpenOcean loves software, and loves it even more when exceptional companies such as MindsDB come to our attention!
