RTF

At the start of 2025 SPC began some focused research into the application of LLM (AI) augmentation of enterprise data systems with use of RAG, context refinement and knowledge graph techniques.

Our aim is to accumulate foundational knowledge about the practical reality of adopting fledgling tools and services for their integration with live information resources, API and IDE.

Dermot Finn has been leading this exploration and development. He has produced good documentation to account for all his hard work. This initial timeline presents the process of progression and refinement of objectives in two phases with extensive reporting. 

All 5 phases of the research portfolio are now documented

We began with an investigation into self-hosted LLM solutions and built several test servers with integrated GPU that utilise Nvidia, CODA Toolkit and Ollama framework for open source LLM models. These hardware resources allow for a broad base of software experimentation which is ongoing alongside application development. 

Much of the application code development has been made possible with use of high density commercial AI services which are evolving at rapid rate and cost to the environment that we keep in mind while using.

We are pleased to acknowledge the support of Jake Strickland and all at EXN for their input and encouragement.

For more information on the research reported here or to make contact with us please email rtf@spc.org.