Develop cutting-edge innovative technologies in trustworthy AI, including privacy-preserving analytics, secure & robust analytics, and distributed & federated learning. We mainly accomplish this through national-level and EU-funded projects and other funding opportunities.

Develop innovative products, services and customer (use case)-driven trustworthy AI solutions unlocked through the deep tech knowledge and experience in deep learning, AI, privacy-preserving, secure & robust analytics, and distributed & federated learning. We work with our prospective customers to help them answering the following questions: What should we do with our data? How to employ AI to help us with our business? How to do it given that there is a plethora of available AI models to use, but they may incur inherent biases from the internet data, uploading data may leak private info, and training may be very expensive? We can help in accomplishing this more efficiently (cheaper) and in privacy & security preserving manner. We target businesses of all ranges, from SMEs to large-scale—we will do the math for whatever the GPU power used will pay off in the net effect.

Provide consultancy services to prospective customers to help them assess whether their planned or developed AI-based solutions are trustworthy: Is a LLM incurring biases; is it “hallucinating”? When can you trust the results generated by your AI model? When can these results be used with guaranteed safety and/or ethical guarantees in real decision-making scenarios? Can you export your data to a third party due to privacy or legal constraints, or perhaps you have to develop an on-premise solution? Our mission is to provide technical tools, tests and procedures that will systematically evaluate planned or existing AI-based solutions against privacy, security, biases, and safety issues; to provide diagnoses and detection of possible problematic issues; and to offer fixes to the issues identified.