Bring compute to your data: an architecture deep dive with Noosware
Multi-party AI projects often hit the same wall: the workload needs access to data that cannot be shared or copied into another environment. How do you move the project forward while keeping protected data in place? This unique challenge requires a robust, carefully designed architecture that enables you to get the computation to the protected data, not the other way around.
On Thursday, 19 November, join the CTO of Noosware at the AI Innovation Center for a technical deep dive into the compute-to-data paradigm: an approach that brings computation to the data, while the protected source data stays where it is.
We will trace the full workflow: from sending a signed workload to a participating data environment to returning results and execution provenance (a record of what ran and where).
On this deep dive, we walk the whole flow end-to-end:
a signed workload travels to the data;
results and execution provenance come back — and the architecture behind it:
workload-scoped authorization, governed data buckets, encrypted payloads,
and a shared state every participant can verify independently.
We are also going to discuss a real-world case study: a health project, where medical centers analyze genomics, clinical, and imaging data across borders, and no patient data ever moves.
Expect a practical discussion of the architecture, its constraints, and its trade-offs, with time for questions and conversation over drinks. This session is designed for AI engineers, data architects, infrastructure teams, and technical leaders working with sensitive data across organizations.
Thursday, November 19th 2026, 15:00 - 18:00, AI Innovation Center, High Tech Campus 5. Free entry, drinks included.