On Coordinating LLMs and Platform Knowledge for Software Modernization and New Developments

Aktiviteetti: Kutsuttu akateeminen esitelmä


Recent generative and fine-tuning LLMs services have been widely benchmarked and used for various tasks in software development. These LLMs are very powerful but have different qualities for software development tasks and may not be able to deal with complex development tasks in edge-cloud software modernization and new developments due to their generative capabilities and lack of up-to-date (domain) knowledge. Various other common development tasks related to target platforms, deployment configurations, policies, data regulation, observability, to name just a few are not well supported/studied in these LLMs. The developer, however, usually has access to many types of knowledge related to their target modernized environments and constraints but such knowledge has not been well connected to LLMs. In this work-in-progress, we will outline situations where the gaps due to offering from LLMs can be compensated by Platform Knowledge capturing knowledge about, e.g., software, service and infrastructure catalogs, architectural decision records and code patterns/samples. We propose a
framework for coordinating LLMs services and Platform Knowledge. At the starting point of the framework, we will discuss challenges for achieving the coordination centered around a common set of development tasks, quality of analytics for coordinating LLMs, integration of common (ML) software development services, and methods for updating knowledge.
Aikajakso1 marrask. 2023
PidettyIBM Watson Research Center, Yhdysvallat