Research
Algorithmic graph theory: electrical flows, graph Laplacians and routing, with a strong pull toward
implementation and experiments.
MSc thesis · expected Feb 2027
Understanding Single Electrical Routing for Solving Flow-Based Optimization Problems
University of Vienna, supervised by Prof. Gramoz Goranci
Can one fixed set of edge resistances route every source and sink demand with low congestion, all at the same time?
It is an open question in oblivious routing. My thesis attacks it from both sides: large experiments to see what is true, and proofs for the cases the experiments make visible.
What I have built
- An IRLS-based electrical routing solver in Julia on top of Laplacians.jl, validated against combinatorial max-flow.
- A multicommodity extension, run on grids, expanders and “chimera” graphs on the ALMA cluster.
- Every graph up to nine vertices, enumerated with nauty/geng and compared against an exact multicommodity LP (JuMP + HiGHS).
- Analytic proofs for the extremal cases the enumeration turns up.
Next: larger experiments to separate constant, polylogarithmic and polynomial growth, and closed forms for other symmetric graph families.
Earlier: dynamic graph algorithms
Before the thesis I worked on incremental maximum flow and dynamic approximate matching in the same group. When the approach ran into correctness problems I changed direction, and that experience shaped how carefully I validate ideas now: experiments first, proofs for what survives.
Education
2024 - 2027
MSc Computer Science, Scientific Computing
University of Vienna
Algorithms, parallel and high-performance computing, optimisation.
2020 - 2024
BSc Computer Science
Constructor University Bremen (formerly Jacobs University)
Thesis: Finding Patterns in Lorenz Systems. A C++ simulation searching for inputs that produce specific chaotic behaviour, with Python visualisation.
2014 - 2018
Baccalaureate, Mathematics and Computer Science
Colegiul Național „Mihai Viteazul”, Bucharest
Competitive programming and olympiad preparation.