Mukul Gagrani
Scholar

Mukul Gagrani

Google Scholar ID: ERZXPy4AAAAJ
Qualcomm AI Research
Efficient LLMReinforcement LearningCombinatorial OptimizationStochastic Control
Homepage↗Google Scholar↗
Citations & Impact
All-time
Citations
403
 
H-index
10
 
i10-index
10
 
Publications
20
 
Co-authors
12
list available
Contact
GitHubOpen ↗LinkedInOpen ↗
Publications
5 items
Efficient Training-Free Multi-Token Prediction via Embedding-Space Probing
2026
Cited
0
ConFu: Contemplate the Future for Better Speculative Sampling
2026
Cited
0
Fast Forward: Accelerating LLM Prefill with Predictive FFN Sparsity
2026
Cited
0
VOCABTRIM: Vocabulary Pruning for Efficient Speculative Decoding in LLMs
2025
Cited
0
CAOTE: KV Caching through Attention Output Error based Token Eviction
2025
Cited
0
Resume (English only)
Co-authors
12 total
Ashutosh Nayyar
Ashutosh Nayyar
University of Southern California
Yi Ouyang
Yi Ouyang
Preferred Networks
Co-author 3
Co-author 3
Rahul Jain
Rahul Jain
Professor of ECE and CS, USC and Research Scientist, Google DeepMind
Co-author 5
Co-author 5
Harris Teague
Harris Teague
Qualcomm, Inc.
Aditya Mahajan
Aditya Mahajan
McGill University
Marcos M. Vasconcelos
Marcos M. Vasconcelos
Assistant Professor, Florida State University