Zhihao Jia
Assistant Professor
Builds systems for efficient machine learning, including compilers and serving systems for large language models.
Zhihao Jia's lab at CMU works on efficient training & inference, compilers & programming languages and large language models. 30 papers since 2023, 29 at top venues, mostly ASPLOS and NeurIPS. 12 alumni, most often to Citadel and Meta. Applications close .
At a glance
- papers, 2023–2026
- 8/ year
- at T1 venues
- 97%
- PhD students
- 0
- MS students
- 0
- alumni listed
- 12
Papers by yearfrom the lab homepage, 2023–2026
- 5 papers2026 · T1 5 · collecting
- 7 papers2025 · T1 7
- 11 papers2024 · T1 10
- 7 papers2023 · T1 7
T1 venuesOtherVenues: ASPLOS, NeurIPS, NSDI, OSDI, ICLR
Research focuspaper titles that use each topic's words
- Efficient training & inference11 papers
- Large language models11 papers
- Compilers & programming languages0 papers
- efficient machine0 papers
- builds systems0 papers
Paper list
From the lab homepage.
20265 papers
- EuroSys (T1)AdaServe: Accelerating Multi-SLO LLM Serving with SLO-Customized Speculative Decoding (opens in a new tab)
- MLSys (T1)Event Tensor: A Unified Abstraction for Compiling Dynamic Megakernel
- NSDI (T1)FlexLLM: Token-Level Co-Serving of LLM Inference and Fine-Tuning with SLO Guarantees
- ICML (T1)Less Is More: Fast and Accurate Reasoning with Cross-Head Unified Sparse Attention
- OSDI (T1)MPK: A Compiler and Runtime for Mega-Kernelizing Tensor Programs (opens in a new tab)
20257 papers
- ASPLOS (T1)GraphPipe: Improving Performance and Scalability of DNN Training with Graph Pipeline Parallelism
- ASPLOS (T1)Helix: Serving Large Language Models over Heterogeneous GPUs and Network via Max-Flow
- ICLR (T1)MagicPIG: LSH Sampling for Efficient LLM Generation (opens in a new tab)
- OSDI (T1)Mirage: A Multi-Level Superoptimizer for Tensor Programs
- NeurIPS (T1)SpecReason: Fast and Accurate Inference-Time Compute via Speculative Reasoning
- NeurIPS (T1)SuffixDecoding: Extreme Speculative Decoding for Emerging AI Applications
- ICLR (T1)TidalDecode: Fast and Accurate LLM Decoding with Position Persistent Sparse Attention (opens in a new tab)
202411 papers
- ICML (T1)Accelerating Iterative Retrieval-augmented Language Model Serving with Speculation
- SCAtlas: Hierarchical Partitioning for Quantum Circuit Simulation on GPUs
- NeurIPS (T1)Communication Bounds for the Distributed Experts Problem
- ASPLOS (T1)Optimal Kernel Orchestration for Tensor Programs with Korch
- NSDI (T1)Parcae: Proactive, Liveput-Optimized DNN Training on Preemptible Instances
- ACL (T1)Quantized Side Tuning: Fast and Memory-Efficient Tuning of Quantized Large Language Models
- PACMPL (T1)Quarl: A Learning-Based Quantum Circuit Optimizer
- NeurIPS (T1)Sequoia: Scalable and Robust Speculative Decoding
- NeurIPS (T1)SpecExec: Massively Parallel Speculative Decoding For Interactive LLM Inference on Consumer Devices
- ASPLOS (T1)SpecInfer: Accelerating Generative Large Language Model Serving with Tree-based Speculative Inference and Verification
- ASPLOS (T1)SpotServe: Serving Generative Large Language Models on Preemptible Instances
20237 papers
- NSDI (T1)Bamboo: Making Preemptible Instances Resilient for Affordable Training of Large DNNs (opens in a new tab)
- MLSys (T1)Building Verified Neural Networks for Computer Systems with Ouroboros (opens in a new tab)
- OSDI (T1)EinNet: Optimizing Tensor Programs with Derivation-Based Transformations (opens in a new tab)
- SOSP (T1)Heterogeneity-Aware, Goodput-Optimized ML-Cluster Scheduling
- VLDB (T1)SDPipe: A Semi-Decentralized Framework for Heterogeneity-aware Pipeline-parallel Training
- VLDB (T1)TOD: GPU-accelerated Outlier Detection via Tensor Operations (opens in a new tab)
- NSDI (T1)TopoOpt: Optimizing the Network Topology for Distributed DNN Training (opens in a new tab)
Where alumni went12 listed, 25% to industry
Industry 3Academia 9
Most often to
- Citadel
- Meta
- Octo AI
Also to
- Carnegie Mellon University
- Cornell University
- Stanford University
Apply to CMUFall 2027
CMU SCS Computer Science PhD
- Round
- Single round
- Opens
- GRE
- See program page
- English
- TOEFL / IELTS / Duolingo (F-1/J-1 non-native speakers)
- Fee
- USD 100 (USD 80 by early deadline)
Confirmed on the program page.
Similar labsnearest by research summary
- CNSJTU
Xiaofeng Gao
- Graph mining & networks
- CNTsinghua
Jianyong Wang
- Data management & big data
- CNPeking
Tengjiao Wang
- Computational social science
- CNSJTU
Yanmin Zhu
- Time series, anomaly & spatio-temporal data
- GBCambridge
Dong Ma
- Generative & foundation models
- USStanford
Azalia Mirhoseini
- Accelerators, chips & architecture
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.
CMUcloses