I am an astrophysicist focusing on topics like cosmic plasmas, high-energy astrophysical phenomena, and computational physics. I’ve also done some research on computer sciences, statistics, and machine learning. For my research I use tools such as pen & paper and supercomputers.
Currently, I have a joint position as a Postdoctoral Research Scientist in Columbia University and as a Flatiron Research Fellow in Center for Computational Astrophysics, Flatiron Institute (New York, USA). Previously, I was a Nordita Fellow in Nordic Institute for Theoretical Physics (Stockholm, Sweden). I am also an avid open-source science & software advocate — you can find my research and codes freely available from GitHub.
I have a wide range of interdisciplinary research interests. These reflect some of my latest publications:
- High-energy astrophysics: accretion (accretion disks); compact objects (neutron stars, black holes)
- Plasma physics: collisionless plasma dynamics; turbulence; particle acceleration
- Nuclear physics: equation of state of cold ultra-dense neutron matter
- General relativity: ray tracing
- Statistics: Bayesian inference; Monte Carlo methods
- Computer sciences: high performance computing; parallellization paradigms; machine learning; Julia language
- Mathematics: cellular automata models; topology
Notes on physics
- Relativistic plasma physics
- MHD turbulence
- Digital filtering
- Introduction to neutron stars (my PhD thesis)
Most of these research notes are done in connection to some publication. If you want to cite these, let me know and I’ll point you to a right paper.
- Tuomo Salmi (host for Nordita PhD visit; MSc and BSc thesis co-supervisor)
- Maarja Kruuse (host for Nordita PhD visit)
- John Hope (MSc thesis supervisor)
- Jere Kuuttila (MSc thesis co-supervisor)
- Eemeli Annala (PhD collaboration)
- Pavel Abolmasov (Univ. Turku)
- Axel Brandenburg (Nordita)
- Andrei Beloborodov (Columbia Univ.)
- Tyler Gorda (Univ. Virginia)
- Jari Kajava (Univ. Turku)
- Aleksi Kurkela (CERN)
- Cole Miller (Univ. Maryland)
- Farrukh Nauman (Chalmers U.)
- Pauli Pihajoki (Univ. Helsinki)
- Juri Poutanen (Univ. Turku)
- Lorenzo Sironi (Columbia Univ.)
- Andrew Steiner (Univ. Tennessee & ORNL)
- Valery Suleimanov (Univ. Tubingen)
- Aleksi Vuorinen (Univ. Helsinki)
- Alexandra Veledina (Nordita & Univ. Turku)
Recent publications & blog posts
When an accretion disk around a neutron star touches the star, it can form a so-called spreading layer. In a work led by P. Abolmasov, we designed a new 2D spherical spectral code to simulate the dynamics of these layers. Code is, obviously, calle...
Runko simulation framework is finally available for everybody. Check it out at GitHub!
In a work led by Alexandra Veledina we presented a new model for the transitional millisecond pulsars in terms of wind-heated disks.
Together with Farrukh Nauman we explored helically forced 3D magnetohydrodynamic turbulence simulations with ensemble machine learning techniques.
Our new EOS analysis indicates that neutrons stars actually have two distinct phases of matter inside: normal hadronic and a sizable quark matter core!
My teaching material for an introductory course on the new high-level, high-performance programming language called Julia is freely available here.
Together with Tuomo Salmi and Juri Poutanen we studied the possibility of applying Bayesian inference to accreting millisecond pulsar X-ray light curves.
In the project led by Pauli Pihajoki, we took ray tracing and numerical differential geometry to the next level. Images of black holes, disks, neutron stars? Can do!
We used new nested hierarchical Bayesian modeling framework to constrain radius of the neutron star in 4U 1702-429 to be R=12.4km. The presented method turns out to be one of the most accurate ones there is!
Bender is finally here! Together with Pauli Pihajoki, we developed a ray tracing platform for rapidly rotating oblate neutron stars. And yes, it's open source!
In the work led by Valery Suleimanov, we applied the direct cooling tail method to constrain the atmosphere composition of a neutron star in 1820. Our results show that helium is strongly favored.
In the work led by Jari Kajava, we studied the interface between the neutron star and the accretion disk. This so-called spreading layer seemed not as constant as people have previously assumed!
How do X-ray bursts really cool? Is it a single-index powerlaw or maybe an exponential? We investigated this together with our student Jere Kuuttila.
Here we derived some corrections to our existing cooling tail method. This new method is then applied to a previously measured neutron star, identifying about 500 meter calibration error.
A collection of scientific writing tips for Master students working with their thesis.
Some peculiar X-ray bursts are found to have atmospheres that are made almost completely of heavy metals. A radiatively driven wind might then ejects some of the burning ashes into the interstellar space.
3 hour introductory lecture material into the world of UNIX.
Equation of state of the cold dense matter has remained mystery for decades. In this recent work, we set new constraints to the EoS parameters by using X-ray bursts that show signs of passive cooling.
What happens to the observed spectrum of neutrons stars when the atmosphere is full of heavy metals from the nuclear burning? On this article, published in A&A we test the effects by computing detailed atmosphere models.
Contains some doodlings I did in the world of hydrodynamics. At some point I became interested in how do fluids actually move so I build my own hydro code to find out.
In this work we show how the accretion modifies the cooling of X-ray bursts on top of neutron stars. This will then lead to (at least) two different classes of bursts, that differ considerably.
On this article, published in MNRAS, we show how the accretion can affect the NS mass and radius measurements that are done using the X-ray bursts.
A project I did of Cellular Automata. During my Christmas holidays I did some studies of how self-organizing criticality works. And for that one needs a proper Cellular Automaton code, of course!