EVENT

Event News

Talk on "From Design to Deployment: Agent-Assisted AI/Data Pipeline Construction with Compliance Awareness" by Adela Nedisan Videsjorden

We are pleased to inform you about the upcoming seminar by Adela Nedisan Videsjorden titled:"From Design to Deployment: Agent-Assisted AI/Data Pipeline Construction with Compliance Awareness" Everyone interested is cordially invited to attend!

Title:

From Design to Deployment: Agent-Assisted AI/Data Pipeline Construction with Compliance Awareness

Abstract:

This presentation introduces inLUMEN, a visual workspace for designing, editing, and generating deployment artifacts for AI and data-processing pipelines using Large Language Model-based (LLM) agents. The tool supports interactive pipeline construction, agent-assisted editing and pipeline refinement, and the generation of deployment artifacts such as Dockerfiles and executable workflow files. By combining visual design with agent-assisted editing, inLUMEN helps users iteratively develop pipeline designs. The presentation also briefly discusses how inLUMEN can integrate with external assessment tools. One example is LexAlign, a multi-agent LLM-based compliance engine that transforms regulatory text into executable decision flows and performs automated compliance assessments. Whilst challenges around reliability, traceability and interpretation exist, inLUMEN points toward a practical solution for building AI/data pipelines following compliance-by-design principles.

Bio:

Adela Nedisan Videsjorden

Adela Nedisan Videsjorden works as a Research Scientist at SINTEF Digital in Norway, as a member of the Trustworthy Green IoT Software research group. Adela has completed a bachelor degree in Language Technology and a Master's degree at University of Oslo in programming and system architecture. Her current research area is centred around the development and enhancement of digital twins using a combination of knowledge graphs representations, Large Language Models and multi-agent technology. Other research areas of interest consist of topics such as edge AI, dialog systems and AI for software engineering.