About

Hi all, and welcome to my academic website.

I’m currently unaffiliated. I call this phase of my life undefined, but I take on research projects whenever it’s possible to actually bring them from raw data or an idea to fruition, using my personal laptop or at most Google Colab or similar. By training, I’m a theoretical nuclear physicist, and recently, via an unlikely project (see below), I found myself pulled back toward nuclear structure, machine learning methods in theoretical physics, surrogate models, and eigenvector continuation.

In addition to physics work, I have also worked on machine learning methods, mainly in osteoarchaeology and forensic odontology, in web development as it applies to building and deploying web apps that host ML models, and I also have an interest in 3D graphics that I used to build a plugin for the data curation platform Clowder of the University of Illinois. Outside of research I make generative digital art and run a physics YouTube channel, and I’m always glad to hear from people working on something interesting.

If you’ve got a physics research question or a machine learning project where the data has a real signal to model, I’d like to hear about it. Same goes for collaborations and job inquiries: get in touch.

Beyond research

I enjoy using Three.js for visualizations and 3D web applications, partly for work and partly because I picked it up while working for Clowder. Three.js provided the initial pull toward visual, interactive web and maybe one day game development, and later AI models for producing graphics, images, and the like. That pull is still there, and if you ask me I will tell you more and show you works I print and so on.

I also run a YouTube channel, recently accepted into the YouTube Partner Program, where I initially uploaded videos walking through A-level physics past papers and lessons. I’m currently also attempting uploads related to research papers, with the ambition being to derive results from papers I understand, show and develop code that people can use, etc. The first such attempt came with an eigenvector continuation paper by Frame et al., “Eigenvector Continuation with Subspace Learning” (Physical Review Letters, 2018). Working on the video took me back to second quantization and revived my interest in theoretical physics. YouTube is a very new direction in my life and its development is ongoing. You can ask me about it too.

Independent research

Since 2025 I’ve managed to publish independently, trying to juggle a few projects at once.

With my former PI Efthymia Nikita, we built a machine learning model that estimates sex from archaeological human remains excavated at ancient Dion, Greece. Skeletal assemblages from archaeological sites are often incomplete, so a good part of the work was handling heavy missing data through imputation; the model reaches around 90 percent accuracy. Published as “Skeletal Sex Estimation for Human Remains From Archaeological Contexts: Machine Learning Models Based on Ancient Dion, Greece”, International Journal of Osteoarchaeology (2025).

Separately, with a group of researchers from Switzerland, Brazil, Bosnia and Herzegovina, Lebanon, and Indonesia, I helped build a forensic dentistry model that classifies whether someone has reached the legal age of majority from panoramic dental radiographs, using transfer learning on pretrained Vision Transformer and EfficientNetV2 models, published as “Classifying Legal Age of Majority (≥18 years) from Panoramic Radiographs with Transfer Learning: Benchmarking ViT and EfficientNetV2”, Journal of Forensic and Legal Medicine (2026).

I also spent May and June of 2025 as a Visiting Researcher at the Université Libre de Bruxelles, where I built machine learning models that predict nuclear potential energy surfaces, essentially a map of how a nucleus’s energy changes as it stretches and deforms. They were trained on Hartree-Fock-Bogoliubov calculations, a standard but computationally expensive nuclear structure method, and reached an R squared of 0.98 on nuclei they had never seen during training. Fast, accurate emulators like that are genuinely useful for exploring fission barriers and the half-lives behind the processes that power stars. This direction of research is fascinating to me, so if you have ideas or data, please get in touch.

Earlier career

I spent 2019 to 2024 at the Cyprus Institute across a few roles. As a Computational Scientist (CaSToRC, 2019 to 2022), I co-led a work package on Open Science and FAIR data within NI4OS-Europe, a Horizon 2020 project (2019 to 2023) building national Open Science Cloud infrastructure across 15 EU member states and associated countries, in support of the EU’s European Open Science Cloud (EOSC). That work put me in regular contact with researchers in the humanities, which is what first pulled me toward machine learning for osteoarchaeological data: it led to SexEst and AgeEst, two web apps I built for estimating skeletal sex and age at death from human remains, both published, with SexEst later adopted as a demonstrator service within NI4OS-Europe. Building them also got me into web-based 3D visualization with Three.js and 3DHOP, which I later brought into Clowder, a data curation platform originally developed at NCSA, University of Illinois. I then moved into an Associate Research Scientist role (STARC, 2023 to 2024), teaching A-Level Physics part time at the International School of Paphos alongside it (Aug 2023 to Apr 2024), before a short postdoctoral fellowship back at CaSToRC (Sept to Nov 2024).

Earlier still, I was a Visiting Assistant Professor of Physics at Monmouth College (Jan 2018 to Sept 2019), teaching introductory physics, classical mechanics, electromagnetism, and mathematical methods, and working with department chair Christopher Fasano on neutron activation, returning to a similar line of research I worked on for my undergraduate thesis. Before that, I was a Postdoctoral Research Associate at Yale University (Oct 2016 to Dec 2017), working with Francesco Iachello on alpha clustering in light nuclei, atomic nuclei that behave as if built from smaller alpha-particle subunits. My PhD work was at the University of Notre Dame (Graduate Assistant, Aug 2009 to Sept 2016; PhD 2017), under Mark Caprio, developing natural orbitals for the no-core configuration interaction approach, a numerical method for computing the structure of light nuclei from first principles, published in Physical Review C, alongside a side project on generalized seniority, a simplifying approximation for medium-mass nuclei. I hold a Diploma in Applied Mathematics and Physical Sciences from the National Technical University of Athens, Greece (2009).

A full list of publications is available on my Publications page, and my talks are listed on my Talks page.