Research around agent evaluation, immersive learning, and measurable AI systems.

Abstract: We present an end-to-end, agent-agnostic framework designed to systematically evaluate Large Language Model (LLM) agents. By isolating the execution and evaluation protocols from internal agent implementations, our platform enables reproducible side-by-side benchmarking across diverse cognitive architectures, establishing clean benchmarks and data-driven metrics for future agent developments.

Abstract: This study proposes immersive virtual reality (VR) solutions to elevate the quality of practical education in technical universities. By developing realistic, interactive simulation nodes of expensive and complex laboratory environments, we facilitate hands-on learning, mitigate hardware constraints, and verify an increase in practical retention rates.