This report documents the experimental validation of a Meta-Reinforcement Learning (MetaRL) framework applied to building temperature control under uncertain thermal dynamics. The work was conducted at the SYSLAB Risø Campus, Danish Technical University, within the scope of the ERIGrid 2.0 Lab Access programme. The goal was to investigate how a controller trained in simulation, using an approximate building model, could adapt to real-world conditions through data-driven adaptation. The experimental activities encompassed initial model identification, parameter uncertainty analysis, simulation-based controller training, real-world data collection, and deployment of a learning-based adaptation mechanism. The targeted building, Power Flexhouse 03, was equipped with temperature sensors and actuators connected to a central control platform. Key findings demonstrate that the proposed Meta-RL framework substantially reduces energy consumption without compromising thermal comfort, with the adaptation mechanism effectively correcting initial modeling inaccuracies after only three days of data. This project shows that Meta-RL is a viable, scalable solution for adaptive control in buildings and can serve as a foundation for next-generation intelligent HVAC systems