ESA BIC Bavaria alumnus Erium GmbH is the technological leader in Artificial Intelligence (AI) optimisation with a minimum of data. Today, AI is choosing the right machine parameters for production processes, optimising forecasts or setting prices for maintenance contracts – and it has great potential in many industrial decision-making processes. The high demands of complex processes require custom-made machine learning models. While it is extremely tedious and expensive to employ Data Scientists to build these models, Erium‘ AI solution empowers process experts, engineers and production managers to create the models independently.
Erium founder Dr Theo Steiniger (TS) was interviewed by Regine Heue (RH), Head of Marketing & Communications at ESA BIC Bavaria’s managing company AZO.
RH: Please give us a sneak peek of your product and the team behind it.
TS: Erium is the technology leader in making process optimisation through AI (Artificial Intelligence) with minimum data accessible to everyone. AI in general has great potential in industrial decision-making. Our AI software calculates e.g. optimal machine parameters in production, optimises forecasting predictions and provides sound figures for the pricing of maintenance contracts. The self-explanatory user interface enables all process experts to use the AI software – without any data science knowledge. This makes the potential of AI accessible to everyone.
The Erium team is made up of an interdisciplinary team of astrophysicists, mathematicians, computer scientists, marketers and business economists. Due to the different backgrounds of the individual team members, everyone has a different perspective on the topics at hand. This enables us to constantly question ourselves and ensure that we create the optimal solution for the customer.
RH: What has been your company’s biggest challenge so far?
TS: The ignition spark of Erium was a specific demand of industrial manufacturing companies for machine learning solutions. From this, the default company development points towards a project-oriented consulting business. Hence, the biggest challenge so far was to figure out the common denominator of all our projects, abstract this very core and create a product-based solution for it that has the potential to be scaled up.
RH: Can you explain the three main ingredients of your recipe for success?
TS: Firstly, we always aim to combine our expertise in data science and software engineering with the use-case-specific engineering know-how. We are convinced that you only have a chance to deliver the best solution possible if you are willing to get your hands dirty as a data scientist.
Secondly, when creating a solution like HALerium, we always have our customer in mind. We have seen numerous tools fail because they were not accessible for the customer. This is the reason why we believe in short feedback cycles.
Thirdly, we use state-of-the-art statistics and machine learning algorithms to combine human expert knowledge with the information that is hidden in data. This is not only fulfilling for our data scientists to research and work with. From this, our customers, in particular, gain the crucial bit of extra performance compared to standard methods that have been on the market so far.
RH: What do you expect from ESA BIC Bavaria and how do you think the initiative and its network will help you to kick-start your business case?
TS: Besides the helpful financial support, the frequent feedback on the business model and growth strategy is extremely helpful in getting things right first time. The ESA BIC network is full of experts to whom you would hardly or possibly never gain access without being part of a programme like ESA BIC Bavaria.
RH: What do you think to benefit from the DLR support?
TS: As Erium, the DLR operates at the intersection between high-tech research and real-world application. Hence, we get helpful insights regarding potential fields of application for our technology and product HALerium.
RH: Let’s reach for the stars – what is going to happen in Erium’s future?
TS: In 2020, we will launch HALerium as a software-as-a-service solution to make it accessible for everyone as easily as possible. Based on this, we will establish HALerium as the standard application for process modelling and optimisation with machine learning, without needing to have mastered the data science that happens behind the scenes. Based on our work in the fields of manufacturing, we will unlock the market potential of sales forecasting, risk assessment and much more.
RH: Many thanks Theo for your time and the interesting insights.