I am an Applied Scientist in the Learned Systems Group at Amazon Web Services (AWS), based in NYC, where I research and build LLM-powered data systems.

I received my Ph.D. in Computer Science from the University of Wisconsin-Madison (2025), advised by Prof. Shivaram Venkataraman. My research focused on improving the performance, efficiency and decision-making of data systems and memory-tiering systems via scalable, AI-guided optimizations. In 2025, I was awarded the Landweber NCR Graduate Fellowship in Distributed Systems for my work.

During my Ph.D., I was fortunate to collaborate closely with Microsoft's Gray Systems Lab. Further, I had the chance to intern with leading industry research teams, including the Learned Systems group at Amazon Web Services (2023) and the Data Systems group at Microsoft Research (2021).

Prior to my Ph.D. journey, I obtained a M.Eng. in Electrical and Computer Engineering (ECE) from the University of Thessaly (2018). I completed my undergraduate thesis with the Computer Systems Lab, where I worked on characterizing program behavior and detecting online phases using low-level CPU metrics. During my studies, I also interned at IBM Research-Zurich (2019) and CERN openlab (2017).

Education and Experience

Publications

  1. Striking the Right Chord: Parameter Tuning in Memory Tiering Systems

    Konstantinos Kanellis*, Sujay Yadalam*, Shivaram Venkataraman, Michael Swift

    DIMES @ SOSP'25: Proceedings of the 3rd Workshop on Disruptive Memory Systems

    Paper

  2. ARMS: Adaptive and Robust Memory Tiering System

    Sujay Yadalam*, Konstantinos Kanellis*, Michael Swift, Shivaram Venkataraman

    arXiv (In submission)

    Preprint

  3. From FASTER to F2: Evolving Concurrent Key-Value Store Designs for Large Skewed Workloads

    Konstantinos Kanellis, Badrish Chandramouli, Ted Hart, Shivaram Venkataraman

    PVLDB'25: Proceedings of the VLDB Endowment, Vol. 18

    Paper Code

  4. From Good to Great: Improving Memory Tiering Performance Through Parameter Tuning

    Konstantinos Kanellis*, Sujay Yadalam*, Fanchao Chen, Michael Swift, Shivaram Venkataraman

    arXiv (In submission)

    Preprint

  5. TUNA: Tuning Unstable and Noisy Cloud Applications

    Johannes Freischuetz, Konstantinos Kanellis, Brian Kroth, Shivaram Venkataraman

    EuroSys'25

    Paper Code

  6. Nautilus: A Benchmarking Platform for DBMS Knob Tuning

    Konstantinos Kanellis, Johannes Freischuetz, Shivaram Venkataraman

    DEEM @ SIGMOD'24 – Proceedings of the 8th Workshop on Data Management for End-to-End Machine Learning

    Paper

  7. Performance Roulette: How Cloud Weather Affects ML-Based System Optimization

    Johannes Freischuetz, Konstantinos Kanellis, Brian Kroth, Shivaram Venkataraman

    ML for Systems Workshop @ NeurIPS'23

    Paper

  8. LlamaTune: Sample-Efficient DBMS Configuration Tuning

    Konstantinos Kanellis, Cong Ding, Brian Kroth, Andreas Müller, Carlo Curino, Shivaram Venkataraman

    PVLDB'22: Proceedings of the VLDB Endowment, Vol. 15

    Paper Slides Code

  9. Too Many Knobs to Tune? Towards Faster Database Tuning by Pre-Selecting Important Knobs

    Konstantinos Kanellis, Ramnatthan Alagappan, Shivaram Venkataraman

    HotStorage'20: 12th USENIX Conference on Hot Topics in Storage and File Systems

    Paper Slides

  10. A Programming Model and Runtime System for Approximation-Aware Heterogeneous Computing

    Ioannis Parnassos, Nikolaos Bellas, Nikolaos Katsaros, Nikolaos Patsiatzis, Athanasios Gkaras, Konstantinos Kanellis, Christos D. Antonopoulos, Michalis Spyrou, Manolis Maroudas

    FPL'17: 27th International Conference on Field Programmable Logic and Applications

    Paper

Contact

Best way to reach me is email.

Email: kkanellis.wisc gmail.com

Links: GitHub · LinkedIn · Scholar · X