Neeraja Yadwadkar
Assistant Professor
Builds cloud computing systems for machine learning, using ML for resource management and efficient model serving.
Neeraja Yadwadkar's lab at UT Austin works on efficient training & inference and distributed, cloud & edge systems. 29 papers since 2023, 5 at top venues, mostly CVPR and EMNLP. 5 PhD students and 2 master's students. 3 alumni, most often to Meta and NVIDIA. Applications close .
At a glance
- papers, 2023–2026
- 8/ year
- at T1 venues
- 17%
- PhD students
- 5
- MS students
- 2
- alumni listed
- 3
Papers by yearfrom the lab homepage, 2023–2026
- 8 papers2026 · T1 4 · collecting
- 12 papers2025 · T1 0
- 4 papers2024 · T1 1
- 5 papers2023 · T1 0
T1 venuesOtherVenues: CVPR, EMNLP, ICLR, OSDI, SOSP
Research focuspaper titles that use each topic's words
- Efficient training & inference6 papers
- Distributed, cloud & edge systems6 papers
- builds cloud0 papers
- model serving0 papers
- resource management0 papers
Paper list
From the lab homepage.
20268 papers
- SOSP (T1)Beyond Utilization: Energy-Conscious GPU Sharing for Inference Serving (opens in a new tab)
- ICLR (T1)Composer: A Search Framework for Hybrid Neural Architecture Design (opens in a new tab)
- CVPR (T1)DeDelayed: Deleting remote inference delay via on-device correction (opens in a new tab)
- ACM SIGOPS Operating Systems ReviewElastic Memory Remapping for Multi-tenant LLM Serving (opens in a new tab)
- TACOET: An Energy Efficient Edge Transformer Architecture (opens in a new tab)
- OSDI (T1)Hardware Lifecycle-Aware Power Planning in Commercial Hyperscale Datacenters (opens in a new tab)
- ACM SIGOPS Operating Systems ReviewRethinking LLM Deployment for Intent-Based Serving (opens in a new tab)
- DIMESTierSizer: Counterfactual Reasoning for DRAM Sizing in Tiered Memory Systems (opens in a new tab)
202512 papers
- SoCCALAP: Intent-Based Serverless Computing via Delayed Decision-Making (opens in a new tab)
- Great Lakes Symposium on VLSI (GLSVLSI)CADOSys: Cache Aware Design Space Optimization for Spatial ML Accelerators (opens in a new tab)
- IEEE Data Compression ConferenceLearned Compression for Compressed Learning (opens in a new tab)
- MLArchSysLimiting Network Bandwidth to Unleash Throughput for Serverless systems (opens in a new tab)
- International Conference on Field Programmable Technology (FPT)LL-ViT: Edge Deployable Vision Transformers with Look Up Table Neurons
- SCMaverIQ: Fingerprint-Guided Extrapolation and Fragmentation-Aware Layering for Intent-Based LLM Serving (opens in a new tab)
- IEEE Computer Architecture Letters (CAL)Old is Gold: Optimizing Single-threaded Applications with ExGen-Malloc (opens in a new tab)
- SoCCOneiros: KV Cache Optimization through Parameter Remapping for Multi-tenant LLM Serving (opens in a new tab)
- ACM Transactions on Architecture and Code Optimization (TACO)Performance Implications of Pipelining the Data Transfer in CPU-GPU Heterogeneous Systems (opens in a new tab)
- Workshop on Advancing LLM-Based Multi-Agent Collaboration (WMAC) at AAAISimulating Rumor Spreading in Social Networks using LLM Agents (opens in a new tab)
- HotCarbonThe Utilization Fallacy and the Real Drivers of Carbon-Efficient Inference Serving (opens in a new tab)
- SESAME @ ASPLOS & EuroSys'25Understanding Sidecars in Cloud Orchestration (opens in a new tab)
20244 papers
- IEEE Computer Architecture Letters (CAL)Characterizing Machine Learning-based Runtime Prefetcher Selection (opens in a new tab)
- SCFast and Efficient Scaling for Microservices with SurgeGuard (opens in a new tab)
- IEEE Data Compression ConferenceMachine Perceptual Quality: Evaluating the Impact of Severe Lossy Compression on Audio and Image Models (opens in a new tab)
- EMNLP (T1)MOSEL: Inference Serving Using Dynamic Modality Selection (opens in a new tab)
20235 papers
- MLArchSysLightweight ML-based Runtime Prefetcher Selection on Many-core Platforms (opens in a new tab)
- HotOSNextgen-malloc: Giving memory allocator its own room in the house (opens in a new tab)
- MLArchSysOnline Learning for Right-Sizing Serverless Functions (opens in a new tab)
- IISWCPerformance Implications of Async Memcpy and UVM: A Tale of Two Data Transfer Modes (opens in a new tab)
- HotInfraSidecars on the Central Lane (opens in a new tab)
Where alumni went3 listed, 100% to industry
Industry 2Academia 0
Most often to
- Meta
- NVIDIA
Apply to UT AustinFall 2027
Computer Science PhD
- Round
- Single round (Dec 15)
- Opens
- GRE
- See program page
- English
- See program page
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- CNSJTU
Chao Li
- Accelerators, chips & architecture
- CNTsinghua
Yu Wang
- Accelerators, chips & architecture
- USUCSD
Hao Zhang
- Efficient training & inference
- USUSC
Viktor K. Prasanna
- Accelerators, chips & architecture
- USUSC
Konstantinos Psounis
- Efficient training & inference
- USUCLA
Harry Xu
- Distributed, cloud & edge systems
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
UT Austincloses