Internship - Machine Learning Research Engineer (Berlin)

This is a full-time, in-person internship in Berlin focused on advancing search quality through models, data, and tooling. You will train and optimize large-scale deep learning models in PyTorch, using distributed training (PyTorch Distributed, DeepSpeed, FSDP) and hardware acceleration for retrieval and ranking tasks. Research representation learning, including contrastive learning, multilingual and multimodal modeling for search and retrieval. Build and optimize Retrieval-Augmented Generation (RAG) pipelines for grounding and answer generation. Contribute to evaluating and improving search quality using robust metrics and principled evaluation. Ideal candidates bring strong PyTorch experience, distributed training know-how, and interest in dense and sparse representations, cross-lingual alignment, and publication-ready research.