![]() |
Sergey Pupyrevspupyrev @ gmail
algorithms
graph drawing
graph theory
compilers
|
I am a scientist working on applied research problems. I am interested in combinatorial optimization, algorithmic graph theory, computational geometry, and information visualization. My current work focuses on developing algorithmic solutions for improving the efficiency of infrastructure and compiler optimizations. Prior to joining industry, I spent several years working on algorithmic graph theory and computational geometry at the University of Arizona (Tucson, USA), Microsoft Research (Redmond, USA), and the Ural State University (Ekaterinburg, Russia), where I received a PhD in Computer Science. I am particularly excited about linear layouts of graphs, which have applications in graph drawing, data compression, compiler optimization, distributed computation, and many other areas. |
News
- June 2023 - A paper on Optimizing Function Layout for Mobile Applications on improving mobile compilers in LCTES'23
- May 2023 - A new paper on Robust and fair work allocation in KDD'23
- April 2023 - My latest long run at 127th Boston Marathon. What an amazing event!
- April 2023 - A new paper on Linear Layouts of Bipartite Planar Graphs accepted at WADS'23
- December 2022 - A paper on Lazy Queue Layouts of Posets published at Algorithmica
- December 2022 - A new paper, Approximating the Minimum Logarithmic Arrangement Problem, accepted at ISAAC'22
- September 2022 - A new pre-print on Minimum Coverage Instrumentation which significantly simplifies code profiling
- August 2022 - A new paper Queue Layouts of Two-Dimensional Posets accepted at GD'22
- August 2022 - A new paper on The Mixed Page Number of Graphs accepted at Theoretical Computer Science
- May 2022 - I am on the Program Committee of The 30th International Symposium on Graph Drawing and Network Visualization
- December 2021 - Ran the first 3-hour marathon: (3:00:51)
- October 2021 - A new paper Profile Inference Revisited accepted at POPL. Here is a teaser of the talk
- April 2020 - A paper resolving a thirty-year-old problem on book embeddings: Four Pages Are Indeed Necessary for Planar Graphs