At the KI4KMU project, I work on accelerating AI adoption for Swiss SMEs by translating advanced AI research into practical, measurable business outcomes. My work focuses on engineering multi-agent AI workflows using Reinforcement Learning and Hybrid-RAG architectures, combining Vector Databases with Knowledge Graphs to improve retrieval quality for Small Language Models (SLMs). I also containerize and deploy AI pipelines with Docker to ensure reproducibility and scalable experimentation. Through this approach, we achieve highly cost-effective AI performance using models up to 97× smaller than traditional commercial LLMs.

Developed internal tools for managing university–company agreements on Microsoft SharePoint using Microsoft Power Apps and Microsoft Power Automate. Also worked with Microsoft Microsoft Azure services to support the deployment, integration, and maintenance of internal applications.