Nils Reimers

Also known as Nils Fabian Reimers

Nils Reimers is a natural language processing researcher who works on text embeddings and search, best known for Sentence-BERT and the sentence-transformers library. After a PhD at the UKP Lab of TU Darmstadt in 2018, he and Iryna Gurevych published Sentence-BERT in 2019, and he coauthored the BEIR benchmark in 2021. At Hugging Face from 2021 to 2022 he built the Neural Search research team, and in 2022 he coauthored the MTEB embedding benchmark and joined Cohere. He co-wrote Cohere's posts on Wikipedia article embeddings in 2023 and on the Rerank 3 model in 2024. As of October 2026, he is VP of AI Search at Cohere, where his team develops search methods built on large transformer networks.

Now at: Cohere
Badges: Bridge

Timeline

2007–2014

Degrees in Oldenburg and Darmstadt

He earns bachelor's degrees in computer science and mathematics at the University of Oldenburg in 2011. In 2014, he graduates from TU Darmstadt with a Master of Science in IT-Security, focusing on algebraic cryptography.
2014–2018

A PhD in the UKP Lab

He works as a doctoral researcher in the UKP Lab at TU Darmstadt. In 2018, he finishes a PhD summa cum laude in deep learning and natural language processing, supervised by Iryna Gurevych.
2019–2025

Sentence-BERT and its library

As a postdoctoral researcher at TU Darmstadt, he coauthors Sentence-BERT with Iryna Gurevych in 2019; the method cuts the effort of finding the most similar sentence pair from 65 hours with BERT to about 5 seconds. Sentence Transformers, the library he started in 2019, moves from the UKP Lab to Hugging Face in October 2025.
2021

The BEIR benchmark

He coauthors BEIR in 2021 with Nandan Thakur, Andreas Rücklé, Abhishek Srivastava and Iryna Gurevych. The benchmark gathers 18 public datasets for zero-shot evaluation of information retrieval models, and it is accepted at the NeurIPS 2021 Dataset and Benchmark Track.
2021–2022

Neural Search at Hugging Face

He joins Hugging Face as a research scientist and builds its Neural Search research team from scratch. The team develops methods that use neural search without labeled training data, and he is the creator and core maintainer of sentence-transformers.
2022

The MTEB benchmark

In October 2022, he coauthors the Massive Text Embedding Benchmark with Niklas Muennighoff, Nouamane Tazi and Loïc Magne. MTEB spans 8 embedding tasks, 58 datasets and 112 languages, and the authors benchmark 33 models on it.
2022–2024

Embeddings and Rerank at Cohere

He joins Cohere in 2022 as Director of Machine Learning, leading the team that develops and trains large language models for semantic search and few-shot text classification. In April 2023, he co-writes Cohere's blog post The Embedding Archives with Jay Alammar, and in April 2024 he co-writes the post introducing Rerank 3 with Sylvie Shi.
As of October 2026

VP of AI Search at Cohere

He is VP of AI Search at Cohere, where his team develops search methods with large transformer networks, including dense and sparse embeddings, neural re-ranking and query generation. In October 2026, Cohere lists him among the key contributors to Rerank 5.