From Insight to Enterprise
Driving the future of intelligent systems through entrepreneurial action and professional services. Explore our portfolio of startup spinoffs, community contributions, and strategic advisory.
SOOFI · RESEARCH CONSORTIUM
Building sovereign AI for Europe
31.6B
total parameters
3.2B
active parameters
Made in Germany
27 trillion tokens
National consortium
multi-institution collaboration
WHAT IS SOOFI-S
Soofi-S is a compact 30-billion-parameter model specifically designed for locally running AI agents. It is intended to run on a powerful PC or Mac with a single consumer GPU, handling tasks such as tool and function calling, multi-step workflows, coding, file processing, and personal assistance. Its primary purpose is to enable powerful AI agents to run directly on users’ own devices, including offline and with sensitive personal data that does not necessarily need to be transferred to the cloud.
DATEXIS’S ROLE
The role of the BHT team at DATEXIS is mainly on enabling complex reasoning. Reasoning is essential because science and other complex business tasks rarely work with perfect, standardized datasets. In their daily routine, users must combine incomplete evidence from different sources, judge the strengths and weaknesses of different methods, and continuously revise their hypotheses as new evidence emerges. AI agents can mimic this process by using reasoning to choose tools, evaluate results, criticize hypotheses, and decide what to investigate next. In this sense, reasoning enables AI to reproduce the iterative and uncertain nature of real scientific discovery.
STARTUP SPINOFFS
Ventures founded out of the lab
Pyramics
Founded 2016
Acquired
Pyramics founding members Thomas Fehn, Philip Hempel, John Radmer and Marc-Andre Scheu developed an optical sensor and a platform for analysing the buying behavior of customers at the Point-of-Sale. Alexander Löser is mentoring the team.
Biometric characteristics of the sensor provide an estimation of age, sex, emotions and the line of sight of people in the detection zone. This anonymous metadata cannot be attributed to any individual. A storage of images which may reveal privacy related data does not take place. Companies utilize signals from the sensor for improving their market research in retail or out-of-home media industry. Specific applications include: age and gender-based customer counting, rule mining from cash transactions + sensor data or advertising effectiveness measurements.
The team receives funding for one year from the BMWi program Exist Seed.
Qualifiction
Founded 2017
ACTIVE
The founding team “QualiFiction” is developing a software for predicting bestsellers. “The cloud solution serves publishers to no longer miss the next “Harry Potter” and to better control the marketing budget”, Dr. Ralf Winkler and Gesa Schöning explain (right to left). Prof. Dr. Alexander Löser (L) is the mentor of the team, which is currently supported by the BMWi with the Exist-Gründerstipendium.
Approximately 90,000 books are published annually on the German market, but the fewest succeed in becoming a bestseller. But what exactly makes a book successful? With the software “Bestseller-DNA”, the QualiFiction team decodes the patterns behind the best-selling books. In doing so, they solve the problems of the publishing houses in order to find economic successes and to better deal with trend topics and target groups. This is done by means of machine learning based on data science: Existing bestsellers are evaluated by the software and predictive findings from different analysis regarding topic, sentiment, entities are combined to form a “bestseller score”. The success of the procedure was already tested by a mathematically related method at the University of Stanford. The scientists were able to predict bestsellers – with a hit rate of 80%! The team has already won a publishing house as a cooperation partner.
The team receives funding from the BMWi program Exist-Gründerstipendium.
BEEZ Data
Founded 2018
Alumni
BEEZ Data develops a platform to support the whole process around funding for nonprofits. The platform allows to standardize and automate the search of and application for funding by clubs and other nonprofits. Further, it enables the simple evaluation of funding applications and the tracking of funded projects for foundations, government agencies, and other funding bodies through dashboards. The BEEZ Data team is made up of Peter Plötner, Rudolf Schmidt, Elisa Molena, Antje Grahl, and Prof. Alexander Löser as mentor.
The numerous possibilities for funding and their individual application processes overwhelm most nonprofits, which need to focus on core competencies with their limited resources. BEEZ Data enables the search of and application for funding in one place. Foundations, government agencies, and other funding bodies face the challenge to select projects with the most impact from numerous applications and subsequently verify the success of those projects. BEEZ Data offers an individualized visual overview of application and project tracking data. Our team is supported through the Berlin Startup Stipendium by the ESF and the Senatsverwaltung für Wirtschaft, Energie und Betriebe.
The team receives funding from Berlin Startup Stipendium and Senatsverwaltung für Wirtschaft, Energie und Betriebe.
PARTNERS & FUNDERS