Nemotron-Math: Efficient Long-Context Distillation of Mathematical Reasoning from Multi-Mode Supervision

Shubham Toshniwal is a senior research scientist at Databricks Mosaic. His homepage and CV describe prior work at NVIDIA and FAIR, with research centered on synthetic data, reinforcement learning, agentic reasoning, and mathematical reasoning for language models.

George William Armstrong's OpenReview profile lists him as a researcher at NVIDIA following doctoral study in bioinformatics and systems biology at the University of California San Diego.

Wei Du works on large language models and NLP in NVIDIA's NeMo team. Public sources describe him as a NeMo team researcher focused on LLM training, deployment, and mathematical reasoning, and note a Ph.D. in Computer Science from the University of Arkansas.

Sadegh Mahdavi's homepage identifies him as a PhD candidate in Electrical and Computer Engineering at the University of British Columbia and a research intern at NVIDIA working on mathematical reasoning for Nemotron models. arXiv public sources also list him as a coauthor of Nemotron-Math.

NVIDIA's public author page identifies Igor Gitman as an applied scientist focused on improving the mathematical reasoning abilities of large language models and as a maintainer of NeMo-Skills.

A 2023 keynote bio describes Branislav Kisacanin as a scientist at Nvidia, Chief Scientist of the Institute for Artificial Intelligence Research and Development of Serbia, and a part-time professor at the University of Novi Sad, working on visual perception for autonomous driving.

An official NVIDIA blog post names Ivan Moshkov as part of the NVIDIA team that won the AI Mathematical Olympiad's AIMO-2 competition.

Canonical link

Toshniwal, Shubham

Armstrong, George

Du, Wei

Mahdavi, Sadegh

Gitman, Igor

Kisacanin, Branislav

Moshkov, Ivan