Huzaifa Shaaban Kabakibo

PhD

Email
huzaifa@mail.uni-paderborn.de
Phone
+49 5251 60-1757
Office
O3.164
Address
Pohlweg 51, 33098 Paderborn, Germany
Web
huzifa1.github.io

I am a PhD student in Computer Science at Paderborn University, working in the Computer Networks Group on large language model (LLM) serving systems and AI infrastructure. My research focuses on making LLM inference more efficient and scalable, with particular interest in serving-system optimization, scheduling, autoscaling, GPU sharing, and serverless inference.

Previously, I developed MoRAGBench, a benchmarking framework for Retrieval-Augmented Generation (RAG) pipelines on mobile and Android devices, and worked on efficient LLM inference for resource-constrained edge hardware.

Publications

2026

  • MLSys 2026
    Breaking the Ice: Analyzing Cold Start Latency in vLLM
    Huzaifa Shaaban Kabakibo, Animesh Trivedi, Lin Wang
    Annual Conference on Machine Learning and Systems (MLSys), 2026
  • arXiv 2026
    SelectInfer: Selective Neuron Loading and Computation for On-Device LLMs
    Huzaifa Shaaban Kabakibo, Eric Schniedermeyer, Artem Burchanow, Lin Wang
    arXiv, 2026