Ongoing Research
Efficient LLM Serving Systems
Current Collaborators: ETRI, Microsoft Research Redmond, Samsung Research, KAIST, Samsung DS
- SLO-aware, adaptive KV cache management
- Generalized KV cache reuse
- Accuracy-preserving long-context pruning
- SLO-aware failure recovery for LLMs
- Systems for agentic AI
Continual and On-Device Learning
Current Collaborators: Palantir, SNU, UIUC
- Low-latency, high-accuracy on-device learning
- Adaptive, resource-efficient continual learning
Large-Scale Distributed Training
Current Collaborators: Samsung Research, KAIST, UC Merced, USC
- Learning-based planning for heterogeneous, geo-distributed training
- Fast distributed training on heterogeneous accelerators
Fast and Scalable Big Data Analytics
Current Collaborators: Amazon, Samsung Electronics, SNU
- Efficient caching for iterative analytics
- Data preprocessing for scalable ML pipelines