Scholar
Mostofa Patwary
Google Scholar ID: 0rt4tbMAAAAJ
Director, Applied Deep Learning Research, NVIDIA
Natural Language Processing
Large Scale Deep Learning
High Performance Computing
Parallel
Follow
Homepage
↗
Google Scholar
↗
Citations & Impact
All-time
Citations
10,446
H-index
35
i10-index
57
Publications
20
Co-authors
13
list available
Contact
GitHub
Open ↗
Publications
12 items
Before They Can Solve: Predicting Post-Training Coding-Agent Performance from Base Models
2026
Cited
0
Nemotron-TwoTower: Diffusion Language Modeling with Pretrained Autoregressive Context
2026
Cited
0
LatentMoE: Toward Optimal Accuracy per FLOP and Parameter in Mixture of Experts
2026
Cited
1
Nemotron Elastic: Towards Efficient Many-in-One Reasoning LLMs
2025
Cited
0
NVIDIA Nemotron Nano V2 VL
2025
Cited
0
Multi-Agent Evolve: LLM Self-Improve through Co-evolution
2025
Cited
0
RLP: Reinforcement as a Pretraining Objective
2025
Cited
0
Nemotron-CC-Math: A 133 Billion-Token-Scale High Quality Math Pretraining Dataset
2025
Cited
0
Load more
Resume
Academic Achievements
Published numerous high-impact papers at venues including NeurIPS 2022, ACL 2021, EMNLP 2020, EACL 2023, etc.
Contributed to Megatron-Turing NLG 530B, the world’s largest and most powerful generative language model at the time
Proposed the Minitron approach for LLM pruning and distillation (ArXiv 2024)
Contributed to StarCoder 2 and The Stack v2 (ArXiv 2024)
Paper 'Scaling Language Model Training to a Trillion Parameters Using Megatron' received Best Student Paper Award at SC 2021
Megatron-LM paper has ~300 citations
Co-authors
7 total
Mohammad Shoeybi
Senior Director of Applied Research at NVIDIA
Bryan Catanzaro
NVIDIA
Alok Choudhary
Professor Northwestern University
Pradeep Dubey
Intel Corporation
Jared Casper
Research Scientist, NVIDIA
Assefaw Gebremedhin
Washington State University, School of EECS
Ryan P. Adams
Princeton University