In the AI Startup of the Week, the editorial staff of ai.nl is featuring promising AI startups, their innovations, solutions and challenges. In this fifth episode, we are taking a look at the Netherlands-based UbiOps, a startup making it easier for businesses to turn their model or script into a live service.
One of the iconic product demos of all time is the late Steve Jobs introducing Macintosh in 1984. That instance when he inserts a floppy disk and the Mac comes to life is a really good example of input becoming output. In today’s world though, a number of businesses and data scientists struggle to bring their model or script to life, or in other words, turn them into a live service.
First-hand experience to the challenge
UbiOps, a Dutch startup based in the Hague was founded by Yannick Maltha and Victor Pereboom, to solve this very challenge. The startup offers a flexible, powerful, scalable, and modern backend for data-driven applications. With the world embracing artificial intelligence and data becoming central to everything that individuals and businesses do, the importance of data-driven applications cannot be stated in mere words.
However, not every business organisation or individual is able to realise the potential of data. UbiOps wants to help anyone looking to bring their models or scripts to life without the need for any advanced cloud or DevOps skills.
The idea originates from first-hand knowledge of the challenge associated with turning an algorithm into a scalable application. After experiencing this challenge, Yannick and Victor began building tools to make it easier for anyone to deploy, serve and manage the lifecycle of an AI program in a single, easy-to-use environment. While it started with analytics, UbiOps is now turning the AI vision of many into reality.
Bringing AI systems from lab to production
One of the common phrases being used in relation to artificial intelligence is the fact that AI applications are no longer confined to research labs. However, it is equally true that there are a number of businesses still finding it difficult to realise the true potential of their data and bring their AI ideas to production.
UbiOps co-founders believe that the gap in tech skills and lack of understanding between software teams and data science teams are some of the reasons. They also see data quality and availability as one of the limiting factors for the effective deployment of AI. “A good data collection and processing infrastructure should become even more of a boardroom priority. Especially with a long-term vision behind it,” Victor Pereboom said at the World Summit AI.
Lastly, Pereboom says there is a need for shared building blocks for AI systems and democratisation of the development of algorithms. UbiOps understands that the primary challenge facing data scientists right now is running their AI reliably in production and it solves this via its feature-rich deployment and serving layer for data science code, models, and scripts.
The startup also takes care of the containerisation of your code, thus offering an option to deploy it as a microservice, request handling, and automatic scaling. The suite from UbiOps also extends to offer features and services that help create dataflows, version management, job scheduling, monitoring, security, and governance.
Building low-code app powered by AI
Low-code is the latest fixation in the world of computing, offering the promise of a field where anyone can develop and deploy codes. The rise of low-code programs has led some to argue that the era of coding is ending but to get there, we need the help of startups like UbiOps. While low-code or no-code platforms offer a lot of promise, they are not an ideal match for AI use cases on their own.
UbiOps makes it possible for a low-code platform development platform like Mendix to work with AI. Mendix makes it easier to develop web and mobile applications in a fraction of the time using pre-built elements and components. However, low-code platforms don’t offer AI functionality and instead push users towards integrations with other tools to realise true potential.
A low-code platform is essentially built around APIs (application programming interface) and the biggest task here is to make your AI call that API. The logical choice here is to make your AI code reside behind an API, which makes it easier to call and set up the connection. The challenge, however, is to get your code behind an API and host it somewhere.
UbiOps makes it easy since it is a platform designed to service data science code and sets up an API automatically. In this article, UbiOps has an explanation of how a low-code app powered with AI can be built simply using Mendix (low-code platform) and UbiOps (the AI backend).
UbiOps and major use cases
The premise of UbiOps is arguably the simplest in the AI field. It allows data scientists and AI-fuelled organisations to run their data science code live in the cloud. There are several use cases where data science can transform and enhance existing industries, like mobility.
Siemens Mobility has partnered UbiOps to offer intelligent mobility applications within urban, interurban and freight transportation. The subsidiary of the German multinational is using UbiOps to deploy, manage, and govern data science and machine learning models at scale. The two companies are jointly focussing on speeding up the development of data science applications for “rolling stock and mobility infrastructure.”
It is not just mobility where UbiOps is making deployment of data science and machine learning models possible at scale, the startup is also helping DuckDuckGoose with their deep fake detection software. Called DeepDetector, the program not only detects deep fakes but also explains its actions. UbiOps helps DuckDuckGoose with building, maintaining, and hosting its image-processing application.
AI and its impact
As a company that helps data scientists solve their challenges with deploying AI applications, UbiOps is in a unique position and its CTO Victor Pereboom thinks that understanding climate dynamics and weather patterns would be one of the major problems that AI can solve. Pereboom says AI will be able to analyse and understand local weather patterns and climate effects in vulnerable areas and UbiOps could be the platform for deployment of such an AI application.
He also sees AI bringing a major revolution in the field of medicine and healthcare, especially in the form of drug discovery, image analysis, and personalised care plans. As a startup that helps data scientists and businesses embrace AI, the startup is acutely aware of the “lack of transparency”.
The startup wants to be clear about where AI is used and what happens with the personal data of people. “We should strive for a system where consumers, and everyone actually, has more ownership on their data and insight in how it is used,” Pereboom says.