Nishil Talati
Assistant Professor
Studies computer architecture and systems for AI, including efficient agentic AI, multimodal models and data analytics.
Nishil Talati's lab at UIUC works on accelerators, chips & architecture, efficient training & inference and data management & big data. 22 papers since 2023. 5 PhD students and 2 master's students. 11 alumni, most often to Meta and Amazon. Applications close .
At a glance
- papers, 2023–2026
- 6/ year
- at T1 venues
- 0%
- PhD students
- 5
- MS students
- 2
- alumni listed
- 11
Papers by yearfrom the lab homepage, 2023–2026
- 5 papers2026 · T1 0 · collecting
- 6 papers2025 · T1 0
- 7 papers2024 · T1 0
- 4 papers2023 · T1 0
T1 venuesOther
Research focuspaper titles that use each topic's words
- Efficient training & inference3 papers
- Data management & big data2 papers
- Accelerators, chips & architecture0 papers
- multimodal models0 papers
- computer architecture0 papers
Paper list
From the lab homepage.
20265 papers
- the IEEE International Symposium on Workload Characterization IISWC .[PDF] [Industry Write-up]Agentic AI Workload Characterization
- International Conference on Very Large Data Bases VLDB .[PDF]Mayura: Expoiting Similarities in Motifs for Temporal Co-Mining
- Conference on Architectural Support for Programming Languages and Operating Systems ASPLOS . [PDF] [Code]MoDM: Efficient Serving for Image Generation via Mixture-of-Diffusion Models
- ACM International Conference on High-Performance Parallel and Distributed Computing HPDC .[PDF]MoE-Lens: Towards the Hardware Limit of High-Throughput MoE LLM Serving Under Resource Constraints
- International Conference on Very Large Data Bases VLDB .[PDF]TIMEST: Temporal Information Motif Estimator Using Sampling Trees
20256 papers
- IEEE/ACM International Symposium on Computer Architecture ISCA .[PDF][Code]DX100: Programmable Data Access Accelerator for Indirection
- the IEEE International Symposium on High-Performance Computer Architecture HPCA .[PDF] [Code]Multi-Dimensional Vector ISA Extension for Mobile In-Cache Computing
- IEEE/ACM International Symposium on Computer Architecture ISCA .[PDF]NMP-PaK: Near-Memory Processing Acceleration for Scalable De Novo Genome Assembly
- the IEEE International Symposium on High-Performance Computer Architecture HPCA .[PDF] [Code]Palermo: Improving the Performance of Oblivious Memory using Protocol-Hardware Co-Design
- International Conference on Very Large Data Bases VLDB .[PDF] [Code]Vortex: Overcoming Memory Capacity Limitations in GPU-Accelerated Large-Scale Data Analytics
- the IEEE International Symposium on Workload Characterization IISWC .[PDF]ZKProphet: Understanding Performance of Zero-Knowledge Proofs on GPUs
20247 papers
- DoE ASCR Energy Efficient Computing for Science Workshop .[PDF]A Deep Co-Design Approach: Systems, Architectures, and Device Cross-Cutting Research for Energy Efficiency
- International Conference on Data Mining ICDM .[PDF] [Code]Accurate and Fast Estimation of Temporal Motifs using Path Sampling
- International Conference on Very Large Databases VLDB .[PDF] [Code] [Supplementary Material]Demystifying Graph Sparsification Algorithms in Graph Properties Preservation
- International Conference on Very Large Databases VLDB .[PDF] [Code]Everest: GPU-Accelerated System for Mining Temporal Motifs
- International Symposium on Computer Architecture ISCA .[PDF]QUETZAL: Vector Acceleration Framework for Modern Genome Sequence Analysis Algorithms
- DoE ASCR Energy Efficient Computing for Science Workshop .[PDF]Toward Cross-Layer Energy Optimizations in AI Systems
- IEEE International Symposium on Workload Characterization IISWC .[PDF] [Code]Understanding The Performance and Estimating The Cost Of LLM Fine-Tuning
20234 papers
- Conference on Architectural Support for Programming Languages and Operating Systems ASPLOS .[PDF] [Code] [Slides] [Lightning Talk]GRACE: A Scalable Graph-Based Approach To Accelerating Recommendation Model Inference
- Design, Automation, and Test in Europe DATE . (Best Paper Honorable Mention)[PDF]PEDAL: A Power Efficient GCN Accelerator with Multiple DAtafLows
- ACM/IEEE International Symposium on Low Power Electronics and Design ISLPED .[PDF]RecPIM: A PIM-Enabled DRAM-RRAM Hybrid Memory System For Recommendation Models
- IEEE International Symposium on Workload Characterization IISWC . (Best Paper Honorable Mention)[PDF] [Code]Vector Processing for Mobile Devices: Benchmark and Analysis
Where alumni went11 listed, 88% to industry
Industry 7Academia 1
Most often to
- Meta
- Amazon
- Apple
Also to
- Carnegie Mellon University
Apply to UIUCFall 2027
UIUC Siebel Computer Science PhD
- Round
- Single round
- Opens
- —
- GRE
- optional
- English
- TOEFL iBT / IELTS
- Fee
- USD 90
Not yet confirmed for this intake.
Similar labsnearest by research summary
- USUT Austin
Aditya Akella
- Accelerators, chips & architecture
- SGNTU
Dmitrii Ustiugov
- Accelerators, chips & architecture
- USGeorgia Tech
Alexey Tumanov
- Efficient training & inference
- CHEPFL
David Atienza Alonso
- Accelerators, chips & architecture
- CNSJTU
Guohao Dai
- Efficient training & inference
- GBOxford
Hamed Haddadi
- Efficient training & inference
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.
UIUCcloses