I am a second-year Artificial Intelligence student at Inland Norway University of Applied Sciences and currently an exchange student at UC Berkeley. My work is focused on practical ML systems: deployment optimization, crash-aware search, LLM serving, benchmarking, and reproducible experimental workflows.

I have built and published projects around constrained ML deployment, crash-aware optimization, and LLM serving autotuning, including SLO-Guard and Constrained-ML-Deployment. These projects reflect my broader interest in making AI systems more reliable, measurable, and honest about what works and what fails.

I am currently building Solon, an autonomous ML research-agent project. The goal is to test whether an AI system can carry out small, benchmarkable research loops: search prior work, propose hypotheses, generate and run experiments, compare against baselines, and produce honest technical write-ups or negative results under human oversight.

I am early in my research career, but I have already spent significant time building systems, running experiments, writing up results, and publishing code and papers publicly. I am applying for support to turn Solon from a promising prototype into a more reliable research system with a concrete, reproducible demonstration.