Founded by a physician and a computer scientist
Bastard Software exists because the hardest parts of precision oncology — clinical judgment and AI/GPU systems engineering — are almost never found in the same room. We put both in the founding team.
Mateusz Bahyrycz
Founder & CEO
Mateusz founded Bastard Software to close the gap between clinical medicine and the low-level compute infrastructure precision oncology actually needs — from SYCL/CUDA compute kernels to on-premise AI inference. He leads the company's technical architecture and the Digital Twin simulation engine.
Rafał Nojek
Co-Founder
Rafał pairs a physician's perspective with deep expertise in machine learning, large language models, and data science. He is the core implementer of Bastard Software's AI stack, particularly OncoKernel's knowledge-graph reasoning and clinical copilot.
Krystian Budek
Co-Founder
Krystian brings a strong background in DevOps, cloud infrastructure, enterprise automation, and software development. He builds automated workflows and integrations across cloud platforms, identity systems, and business applications. He brings his infrastructure and product engineering experience to OncoKernel.
That pairing is not incidental to the product — it is the thesis. Precision oncology software fails when it is built by one discipline in isolation: engineers who don't understand what a tumor board actually needs, or clinicians without the systems background to build AI infrastructure that runs safely inside a hospital network.
We work alongside a practicing oncologist as clinical design partner and first pilot site — supplying the oncology domain validation, and the hospital relationship that turns a demo into a deployment.
Our thesis
Sequencing a tumor is cheap. Interpreting it — safely, explainably, and without the data leaving the hospital — is the hard, valuable problem.
Clinician-led design
Every tool is designed around what a physician actually does with the output, with a practicing oncologist validating the workflow.
Systems-engineering rigor
GPU compute, data contracts, and on-premise deployment are treated as first-class engineering problems, not afterthoughts.
Built for hospitals
Every architectural decision is made to be deployable inside a real hospital network under real regulatory constraints.