Speaker
Mark Fitzsimons
(Microsoft)
Description
As AI becomes embedded in research, the real challenge is not access to models but building them properly: reproducible, auditable, and sovereign by design.
This session explores how research teams are developing AI end to end using Microsoft Fabric, Microsoft Foundry and HPC within trusted, policy‑controlled cloud environments, from governed data and shared semantics through experimentation, large‑scale training and rigorous evaluation. The focus is on real research workflows, where collaborative data preparation, structured model development and high‑performance compute come together to support trust, compliance and scale without forcing centralisation or loss of control.
Treating AI as a scientific instrument, not a black‑box service.
Primary author
Mark Fitzsimons
(Microsoft)
