Bita Darvish Rouhani
Scholar

Bita Darvish Rouhani

Google Scholar ID: oqeh4cYAAAAJ
Distinguished Engineer, NVIDIA
Generative AIAI SupercomputingSystems for AI
Google Scholar↗
Citations & Impact
All-time
Citations
3,655
 
H-index
28
 
i10-index
42
 
Publications
20
 
Co-authors
9
list available
Contact
No contact links provided.
Publications
8 items
Guess-Verify-Refine: Data-Aware Top-K for Sparse-Attention Decoding on Blackwell via Temporal Correlation
2026
Cited
0
SPEED-Bench: A Unified and Diverse Benchmark for Speculative Decoding
2026
Cited
1
Efficient MoE Serving in the Memory-Bound Regime: Balance Activated Experts, Not Tokens
2025
Cited
0
Pretraining Large Language Models with NVFP4
2025
Cited
0
Helix Parallelism: Rethinking Sharding Strategies for Interactive Multi-Million-Token LLM Decoding
2025
Cited
0
Beyond the Buzz: A Pragmatic Take on Inference Disaggregation
2025
Cited
0
Key, Value, Compress: A Systematic Exploration of KV Cache Compression Techniques
2025
Cited
0
ResMoE: Space-efficient Compression of Mixture of Experts LLMs via Residual Restoration
2025
Cited
0
Resume (English only)
Co-authors
9 total
Farinaz Koushanfar
Farinaz Koushanfar
Professor and Siavouche Nemat-Nasser Endowed Chair of ECE, UC San Diego
Eric S Chung
Eric S Chung
VP of AI Computing, NVIDIA
Co-author 3
Co-author 3
Azalia Mirhoseini
Azalia Mirhoseini
Assistant Professor of Computer Science, Stanford - Google DeepMind
Steven K. Reinhardt
Steven K. Reinhardt
AMD
Maxim Naumov
Maxim Naumov
Meta (Director of Engineering & Research)
Co-author 7
Co-author 7
Pradeep Dubey
Pradeep Dubey
Intel Corporation