Konstantinos Kanellis
- CS Ph.D. Student, University of Wisconsin-Madison
- kkanellis cs.wisc.edu
My research interests mostly lie in the area of computer systems, with a focus at the intersection of data systems with machine learning. My latest research aims at developing better database auto-tuning solutions, by applying machine learning techniques. I also work on improving key-value store designs so that they can fully exploit new, fast storage devices.
Before moving to Madison, I received a M.Eng. in Electrical and Computer Engineering from the University of Thessaly, Greece. In my undergraduate thesis, I worked with the Computer Systems Lab, on characterizing program behavior and detecting online phases using low-level metrics from CPUs.
While at UW-Madison, I have interned at Amazon Web Services (2023) and Microsoft Research (2021). Prior to my PhD journey, I also interned at IBM Research - Zurich (2019) and CERN openlab (2017).
News
-
[06.24] Attended SIGMOD'24 in lovely Santiago, and presented Nautilus! 🇨🇱
-
[04.24] Nautilus, a framework thar simplifies the benchmarking and evaluation of configuration knob auto-tuners, was accepted to DEEM Workshop @ SIGMOD! Stay tuned for the open-source release! 🔥
-
[01.24] Passed my Ph.D. Preliminary Examination! 🚀
-
[10.23] Our work on noisy auto-tuning was accepted to ML for Systems Workshop @ NeurIPS! 🔥
-
[05.23] Started my research internship at AWS, Learned Research Group!
-
[05.23] Preprint of our new work (with MSR) on evolving key-value store designs is now available!
-
[09.22] Attended VLDB 2022 in the beautiful city of Sydney, and presented our LlamaTune work! 🇦🇺🦙
-
[06.22] LlamaTune accepted at VLDB 2022! 🔥
-
[12.21] Passed my Ph.D. Qualifying Exams! 😄
-
[05.21] Excited to start my research internship at MSR Redmond in the Data Systems group!
-
[01.21] Started as affiliate research assistant at Microsoft Gray Systems Lab!
-
[06.20] Virtually presented our work on reducing the tuning time of storage systems at HotStorage'20!
Publications
-
Nautilus: A Benchmarking Platform for DBMS Knob Tuning
Konstantinos Kanellis, Johannes Freischuetz, Shivaram Venkataraman DEEM'24 @ SIGMOD – Proceedings of the 8th Workshop on Data Management for End-to-End Machine Learning -
Performance Roulette: How Cloud Weather Affects ML-Based System Optimization
Johannes Freischuetz, Konstantinos Kanellis, Brian Kroth, Shivaram Venkataraman ML for Systems Workshop @ NeurIPS'23 -
F2: Designing a Key-Value Store for Large Skewed Workloads
Konstantinos Kanellis, Badrish Chandramouli, Shivaram Venkataraman ArXiv (in submission) -
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 -
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 -
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
Experience
-
Applied Scientist Intern @ Amazon Web Services – Learned Systems Group, Summer 2023
-
Research Intern @ Microsoft Research – Data Systems Group, Summer 2021
-
Research Intern @ IBM Research – Zurich, Spring 2019
-
Summer Student Intern @ CERN openlab, Summer 2017