Hi!👋, I'm Talia
I'm a first-year Ph.D. student at UT Austin, advised by
Prof. Vijay Chidambaram.
My research interests lie in Distributed Systems and Systems for Machine Learning and Agentic AI.
Before joining UT, I earned my BA/MS in CS from Boston University, where I worked on building efficient training systems for GNN with io_uring under the supervision of
Prof. Vasiliki (Vasia) Kalavri.
News
Research
Projects
-
🤖 Building Adaptive Systems for Agentic Workflow Optimization [On Going]
Developed an adaptive system to optimize agentic workflows for batched query processing, integrating hybrid resource management and dynamic utilization planning to maximize resource efficiency and minimize cost.
-
💾 Distributed Vector DB with CXL for Large-Scale LLM Workloads [On Going]
Designed and implemented distributed vector databases leveraging CXL to handle large-scale datasets and high-throughput LLM inference, improving performance and memory efficiency.
-
🧠Efficient Training Systems for Large-Scale GNNs with io_uring
Built scalable Graph Neural Netwroks sampling systems using Linux's
io_uringfor asynchronous I/O, enabling efficient out-of-core graph training and high-throughput data streaming.
Publications
Workshop
-
RingSampler: GNN sampling on large-scale graphs with io_uring
Qixuan Chen, Yuhang Song, Melissa Martinez, Vasiliki Kalavri
HotStorage '25 [pdf] [slide]
Poster
-
Scaling GNN Sampling on Large-Scale Graphs with io_uring
Qixuan Chen, Yuhang Song, Melissa Martinez, Vasiliki Kalavri
FAST '25 [poster]