Nikos Pagonas

CS PhD Student @ Columbia University

profile.jpeg

Hi! I am Nikos, a second-year PhD student in Computer Science at Columbia University, where I am a member of DAPLab, advised by Prof. Kostis Kaffes.

My research focuses on improving the performance and efficiency of agentic serving. I build systems that exploit workflow structure and runtime characteristics to optimize model routing, scheduling, and resource allocation.

This summer, I am working as a Research Intern at Microsoft Research, supervised by Sharad Agarwal.

In Summer 2025, I was a Student Researcher at Google, where I worked on Cortex, a workflow-aware agentic serving system, supervised by Prof. Arvind Krishnamurthy.

Before joining Columbia, I was a member of the ATLAS research group at Brown University, where I worked on PaSh, a Linux Foundation project, advised by Prof. Nikos Vasilakis.

I received my Master’s degree in Electrical and Computer Engineering from the National Technical University of Athens (NTUA), where I was advised by Prof. Georgios Goumas.

News

Apr, 2026 Our position paper on Agentic Data Environments appeared in the IEEE Data Engineering Bulletin!
Aug, 2025 Cortex made it into the SOSP 2025 Workshop on Systems for Agentic AI!
Aug, 2025 I was awarded the A.G. Leventis Foundation Scholarship!
Jun, 2025 I was awarded the Gerondelis Foundation Scholarship!
May, 2025 I joined Google as a Student Researcher!
Sep, 2024 I was awarded the Columbia Presidential Fellowship!
Sep, 2024 I started my PhD in Computer Science at Columbia University!
Jun, 2024 I received my Master’s degree in Electrical and Computer Engineering from NTUA!
Sep, 2023 I joined Brown University as a Student Researcher!
May, 2023 I started working on the PaSh project!

Selected publications

2026

  1. arXiv 2026
    AgentOpt v0.1 Technical Report: Client-Side Optimization for LLM-Based Agent
    Wenyue Hua, Sripad Karne, Qian Xie, Armaan Agrawal, Nikos Pagonas, Kostis Kaffes, and Tianyi Peng
    arXiv 2026
    arXiv
  2. arXiv 2026
    VineLM: Trie-Based Fine-Grained Control for Agentic Workflows
    Nikos Pagonas, Matthew Lou, Tianyi Peng, Dan Rubenstein, and Kostis Kaffes
    arXiv 2026
    PDF
  3. IEEE DEB 2026
    Agentic Data Environments
    Elaine Ang, Chenxi Huang, Georgios Liargkovas, Jerry Liu, Jinhui Liu, Nikos Pagonas, Charlie Summers, Haonan Wang, Jiakai Xu, Tianle Zhou, Yusen Zhang, Zhou Yu, Zhuo Zhang, Tianyi Peng, Kostis Kaffes, and Eugene Wu
    IEEE Data Engineering Bulletin, March 2026
    PDF

2025

  1. SAA 2025
    Cortex: Workflow-Aware Resource Pooling and Scheduling for Agentic Serving
    Nikos Pagonas, Yeounoh Chung, Kostis Kaffes, and Arvind Krishnamurthy
    SOSP 2025 Workshop on Systems for Agentic AI (SAA 2025)
    PDF