AI That Never Leaves the Vault: Running Python and LLMs Natively Inside SPEs

12.30pm – 12.45pm IST, 13 October 2026 ‐ 15 mins

Room: Wicklow Hall 2 - Level 2

Kerry Kirkham, Senior Sales Engineer, InterSystems

Secure Processing Environments (SPEs) are becoming the standard way to give researchers governed access to sensitive health data — but their value depends on what researchers can actually do inside them.

This use case focuses on embedding AI and analytics directly within the SPE boundary: running Python-based analysis and large language models locally, alongside the data, rather than exporting data to external tools or cloud AI services. This approach preserves data sovereignty and auditability, avoids the compliance risk of moving sensitive data for analysis, and still gives researchers modern capabilities like natural-language querying, summarization, and vector search over health records.