arvid eidenert

Architected and implemented core gameplay systems and seasonal content updates for a Roblox title, maintaining high performance for 15,000+ daily active users. Partnered with animators, 3D modellers, VFX artists, and QA testers to connect technical implementation with visual excellence.

Implemented and evaluated an enterprise AI assistant prototype using Microsoft Copilot Studio and Azure AI Foundry to support IAM workflows via Retrieval-Augmented Generation (RAG). Prepared and analyzed data sources, configured secure conversational logic, applied Design Science Research methodology, and achieved 95.7% identification precision.

Specialized in mathematical and Swedish language tasks for LLM training data, performing high-quality annotation and RLHF work to improve domain accuracy. Evaluated model outputs and reasoning behavior and provided targeted feedback to enhance dataset quality and model performance.

Bachelor of Science in Software Development covering topics such as object-oriented programming, machine learning, data science, distributed computing, cloud computing, and full-stack web development. Bachelor’s thesis implemented and evaluated an enterprise AI assistant prototype for IAM workflows using Microsoft Copilot Studio and Azure AI Foundry with RAG, achieving 95.7% identification precisi