Team
Meet the team
Three co-founders trained in physics, mathematics and statistics. Between us we cover the whole path of an AI system, from the mathematics of the model to the infrastructure it runs on.
Roland Boubela
Co-founder
Vienna
Background
- PhD Medical Physics, Medical University of Vienna
- MSc Statistics, TU Wien
- CTO & Co-Founder at MOSTLY AI (2017–2022)
Focus Areas
- → Architectures for agentic LLM systems
- → Running LLMs on-premises
- → Large-scale compute and data infrastructure for machine learning
- → Building and leading engineering teams
Klaudius Kalcher
Co-founder
Vienna
Background
- PhD Medical Physics, Medical University of Vienna
- MSc Statistics, TU Wien
- Co-Founder & Chief Data Scientist at MOSTLY AI (2017–2022)
Focus Areas
- → Discovery: finding the right AI approach for a given problem
- → Generative model design and synthetic data
- → Ethical AI and compliance: reviewing AI systems against the EU AI Act and GDPR
- → Turning research into shipped products
Andrey Sidorenko
Co-founder
Vienna
Background
- PhD Physics & Mathematics, Moscow State University
- Head of AI Research at MOSTLY AI until 2026; joined as a developer in 2019
- Previously theoretical and experimental physicist; built and controlled systems for near-absolute-zero experiments
Focus Areas
- → Generative AI for tabular data and multimodal systems
- → Differential privacy, synthetic data, anonymization
- → LLM fine-tuning, inference optimization, distributed training
- → Scientific computing: physical modeling, simulation, data analysis
Our Story
The three of us worked together at MOSTLY AI, the first company to use deep learning to generate structured synthetic data. Klaudius and Roland were co-founders in 2017; Andrey joined as a developer and went on to lead its AI research. We took the product from a first prototype to software adopted by several Fortune 100 companies.
What we built was a synthetic data engine: generative deep neural networks developed in-house from early 2017, before the transformer paper was published and years before “generative AI” became a common term. Klaudius designed the models, Roland built the compute infrastructure and the software that turned them into a product we could ship, and Andrey led development of the core AI engine and later took over the research, after Klaudius had moved into other roles in the company. The engine ran in our customers’ own data centers and cloud environments, at companies in banking, healthcare and telecommunications, each serving millions of end customers. Its core has since been released as open source.
We started Hinetra in 2024 because we kept seeing companies hire consultants who couldn't build, or contractors who couldn't design. We do both. Having built a product company ourselves, we help you decide what to build and how, including whether it should be built at all. Then the same people build it and put it into production.