Grigory Grechkin

Research assistant in machine learning and M.Sc. student at RPTU Kaiserslautern-Landau, working on computer vision models that keep working when the context changes.

My research is supervised by Naghmeh Ghanooni and covers synthetic data generation, data selection, feature-space analysis and mitigating spurious correlations. I train and evaluate large-scale self-supervised and contrastive models, integrate diverse datasets into custom pipelines, and design reproducible multi-run experiments.

Alongside that I build and run things end to end: a home server with a dozen self-hosted services behind one sign-in, and a local assistant over my own documents with the model on a GPU it has to share.