Top-10 AI Startups in Central and Eastern Europe

Some people think that “sustainable eating” means shopping exclusively at over-priced chains like Whole Foods but we’re here to shut that myth down with two words – Almazan Kitchen. Who knew that a couple of Serbian guys cooking organic food in a forest with their pet owl could generate nearly 20 million views for a single video? Serbia is just one of many countries found in Central and Eastern Europe (CEE), a region that consists of relatively small countries that gained independence after the collapse of the Soviet Union around 1989. As these countries transformed into democratic states with capitalistic economies only 30 years ago, they had to overcome unique handicaps compared to more developed nations in Western Europe – like no organic grocery stores. Still, most of the region has registered rapid GDP growth over time that was only slowed by the global financial crisis of 2008.

National GDP in USD since 1989 – Credit: World Bank

Some of these countries we’ve looked at in the past, like our articles on the top AI startups in Poland and Estonia. For this article, we queried Crunchbase to look for interesting AI startups in countries like Albania, Bulgaria, Croatia, Czech Republic, Hungary, Romania, Slovakia, and Slovenia. In addition to a rich and diverse cultural heritage that drives a strong tourism industry in the region, many of these countries have become tech outsourcing hubs due to large talent pools, deep technical knowledge, cheap office space, and the fact that you can pay people less than developed markets and get roughly the same value. Lots of tech activity accelerates the growth of startup ecosystems which means venture capitalists are becoming increasingly aware of the region’s potential. Look no further than these AI startups in Central and Eastern Europe with the most published funding according to Crunchbase.

Founded in 2014, Hungarian startup AImotive has raised $47.5 million from investors including Samsung, Bosch, and Nvidia to develop Level 5 autonomy software for self-driving cars. The startup has created a modular software package that uses AI algorithms for perception and motion planning. The software works with many sensors including LiDAR, camera, and ultrasonic. Many kinds of hardware and operating systems are supported including newly developed automotive systems.

AImotive claims their aiDrive platform can already handle highway autopilot and autonomous valet parking in closed confines, while urban autopilot is being introduced to limited city scenarios throughout 2019, and open-air parking is under development.

AImotive holds testing licenses in Hungary, Finland and the states of California and Nevada – Source: Wikipedia

AImotive also offers a simulator to test autonomous driving solutions and an AI accelerator to optimize CPU performance for their software. With offices in Budapest, Silicon Valley, and Japan, the aiDrive solution is already available for licensing if you want to kick the proverbial tires a bit.

Albania may be one of the best countries in the CEE region to visit, and it’s also home to startup Gjirafa which raised $8.7 million to develop the Albanian version of Google. The project started with a search engine in the local language and has expanded to a business directory, e-commerce site, online marketplace, video streaming platform, advertising network, and startup accelerator. A “something for everyone” business model it seems. The founders’ vision was to build and facilitate the internet economy in the Balkans for a market of 12 million people. The first site launched in 2014 and since then Gjirafa has gathered 3 million active monthly users which now surpasses Albania’s population of around 2.87 million people. The startup now has more than 60 employees, some of whom are working on payment processing and AI-powered semantic solutions for the Albanian language.

Founded in 2006, Hungarian startup Starschema has raised $5 million to develop big data analytics for diverse industries. The startup is managing custom projects related to data science, visualization, consulting, and software engineering for a number of sectors. For example, the company helps banks forecast ATM and branch cash needs and manage risks using machine learning. For manufacturing companies, they measure equipment performance, optimize inventory, and manage finances. Healthcare providers can analyze clinical data and provide personalized medicine to clients using Starschema’s algorithms. The startup also offers consulting services related to Tableau, a market leading business intelligence platform. Again, we see a “something for everyone” business model. Starschema is present in Budapest and Arlington, Virginia.

Founded in 2017, Czech Republic startup “Time is Ltd.” has raised $3.4 million to develop a big data platform that analyzes workplace productivity. The app turns internal communications, calendars, Slack, and other cloud systems’ data into visualized insights that help understand patterns of behavior and team dynamics inside a corporate organization. For example, it can show how much time salespeople spend on internal meetings instead of facing clients, or the change in productivity following the introduction of an Agile project management methodology. The tool has native integrations to all leading corporate software suites so it doesn’t require any implementation effort. The startup isn’t alone, and some of their competitors can be found in our previous article on “7 Employee Productivity Tracking Solutions.”

Founded in 2015, Czech Republic startup Trivi has raised $3.2 million to develop an accounting automation and tax optimization app that uses good old Optical Character Recognition (OCR) to handle scanned documents and PDFs, then categorizes the extracted data using machine learning. It also takes care of recurring invoices and payroll processing without human interaction. The algorithms not only manage administrative work, but provide a visual overview of company financials with invoice maturities, payment dates, and cash-flow. Trivi’s accounting team is available for tax consultation and audit management, and can also help in customizing the platform for specific business use cases. Subscriptions start at $17 per month and 18 cents per uploaded document.

Founded in 2007,  Hungarian startup Gravity R&D has raised $3 million to develop a personalization engine for online and offline business models. The engine, called Yusp, automatically learns and analyses the browsing and shopping behavior of users and predicts what they are likely to enjoy based on their similarity to other users. The platform can also recommend relevant content on streaming sites or relevant products on commercial sites and also performs predictive analytics on local purchasing habits.

According to Gravity R&D, 62% of customers have chosen, recommended or paid more for a brand that provides a personalized service, and 86% of customers say personalization plays a role in their purchase decisions. In one of the case studies, French hypermarket chain Cora has achieved 10% more conversions and 7% more revenue after integrating Yusp into its web and mobile commercial platform. Sounds like one step closer to the pinnacle of marketing achievements –  online echo chambers.

Hungarian startup DrugPredict is actually a consortium of Hungarian universities and corporates that has raised $2.8 million from the National Institutes of Health to develop a comprehensive knowledge base of existing FDA-approved small-molecule drugs and other small-molecule compounds where each entry has been modeled for its effects on the human body. Recent developments in pharmacology have been focusing on how drugs affect the body in multiple areas and highly complex ways rather than simply targeting one specific biological factor. Looking at how complex molecular feature sets of drugs correlate with the known drug effects can provide predictive power to reveal the entire effect profiles of drugs.

DrugPredict not only models existing drugs and compounds but can predict the potential effects of user-defined molecules. Such a technology will allow pharmacologists to reposition existing drugs as therapies for new illnesses and develop new compounds with specific effect profiles in mind. According to researchers, the cost of developing a new prescription medicine has increased by 145% over the period 2003-2013 reaching $2.6 billion. This predictive drug discovery method could make the process more streamlined and remove part of the massive development cost. It’s something we touched on in our article on How Computational Chemistry Helps Drug Discovery.

Founded in 2013, Slovakian startup Photoneo has raised $2.4 million to develop 3D machine vision for industrial automation and robotics. The startup offers 3D sensors and associated modeling software that allows accurate mapping of items of any size and shape – a basic requirement for bin picking, where many randomly placed parts need to be recognized and placed uniformly on conveyors or into packaging. Photoneo’s solutions also recognize and sort mixed objects and are natively integrated with major industrial robot manufacturers like ABB, Kuka, and Fanuc.

The many use cases of Photoneo technology – Credit: Photoneo

The company has also built its own autonomous transportation robot on the back of its machine vision technology. The robot is a mobile platform that maps its environment and plans its route safely, transporting up to 100 kg in warehouses, factories, and hotels. Customer testimonials highlight the industrial strength of Photoneo’s hardware, its data quality, and its ease of use as major selling points for the equipment.

Founded in 2013, Bulgarian startup Transmetrics has raised $1.9 million to develop predictive optimization models for cargo transport and logistics companies. The startup uses a three-step approach to optimize logistics. First, they take raw data from transport companies to cleanse and enrich it. Using this standardized dataset, Transmetrics’ algorithms then generate short-term demand forecasts based on historical demand. Finally, forecast results are input into an AI optimizer that translates forecasts into optimal capacity, warehouse, and labor planning. Customers need to start with at least six months of historical data on shipments, tracking information, and vehicle movements, and on-going data is automatically extracted by the platform after launch. The subscription service can reduce linehaul costs by up to 25% and is being used by leading logistics companies worldwide including DHL and DPD.

Founded in 2016, Hungarian startup SignAll has raised $1.7 million to develop an automated real-time sign language translation solution that uses four different cameras to capture body movement, hand shapes, and facial expressions, and then translates them into a chat dialogue. While the deaf person signs toward the cameras, the hearing person’s speech is picked up by voice recognition. Both parties have a screen which displays the whole chat.

SignAll provides and installs its own workstations for clients – Credit: SignAll

The company also offers education terminals for hearing users to practice sign language vocabulary, and for deaf users to learn additional signed languages (there are over 135 different variations over the world). At the moment the platform handles American Sign Language and spoken English, and targets public and private institutions looking to improve their accessibility or workplace diversity. Outside of their headquarters in Budapest, the startup has opened an office in Arlington, Virginia to serve US customers. We did come across one other startup in the U.S. that is doing something similar but now appears to be just a failed crowdfunding campaign. If they let us know otherwise, we’ll update this article.

Conclusion

Whenever we do one of these lists, we’ll inevitably hear from a startup that flew under the radar for whatever reason and thinks it ought to be on the list too. Given the large number of countries in the CEE region, there are probably at least a few more interesting AI startups worth writing about that didn’t make this list. If you’re one of them, drop us a note and let’s have a chat about what you’re getting up to. Bonus points if you have a pet owl roaming around the office.

Worried that AI might steal your job? You should be. Why not get a Masters in Machine Learning from London’s Imperial College, one of the top-10 universities in the world? And you can study online!