Real-time 3D & XR · Scientific visualisation · Astronomy

Alexander Sivitilli

I build real-time 3D and virtual-reality software for large scientific datasets, and I hold a doctorate in how people actually use it. Eight years across GPU rendering, data pipelines, interaction design and digital-dome delivery.

Technical Lead on SpaceDome, an ESA-funded Earth-observation platform, and Leverhulme Postdoctoral Research Associate at Armagh Observatory & Planetarium.

What I build

Three connected lines of work.

Real-time 3D and immersive systems

I build virtual-reality, digital-dome and spatial-computing environments in Unity: ray-marched volume rendering, custom HLSL shaders, native C++ data pipelines and the interaction systems that keep large datasets navigable at framerate.

Human-centred design and evaluation

My doctorate is mixed-methods user research on immersive learning. I run the studies that separate real analytical benefit from novelty, then feed the findings back into the system I built.

Machine learning for scientific data

I design and validate neural networks for asteroid-family identification, and build immersive diagnostics that render model confidence directly in orbital-element space, surfacing disagreements that 2D plots obscure.

Flagship project

iDaVIE: open-source VR for volumetric scientific data.

First author and lead developer of a Unity and SteamVR application for interactive analysis of 3D data cubes, released at v1.0 under LGPL v3, registered in the Astrophysics Source Code Library and published inAstronomy and Computing with twelve authors across five countries. External groups have used it to correct source-detection masks and publish discoveries they credit to the tool.

Toolkit

What I build with.

Where a claim has something public behind it, it links to the code, paper or release that shows it.

Scientific Data & ML

  • Python
  • PyTorch
  • scikit-learn
  • NumPy
  • Neural network design, training and validation
  • Local LLM and speech inference (whisper.cpp, open-weight models)
  • Image processing and custom detection pipelines
  • FITS, HDF5, WCS, VOTable, DICOM and Parquet
  • Large multidimensional and geospatial datasets

About

Building tools, then studying how people use them.

My work sits at the intersection of real-time graphics, immersive systems, scientific software and human interaction. I am interested not only in whether a new interface is technically possible, but in whether it genuinely makes difficult data easier to reason about.

That means combining development with observation and evaluation: creating the system, working with its users and separating productive interaction from novelty. The aim is immersion in service of better scientific questions, not immersion for its own sake.

See how this approach appears across current and previous projects

Contact

Let's make complex data easier to reach.

Currently open to 3D, XR and visualisation engineering roles, in industry or research. I also welcome conversations about immersive scientific visualisation, digital-dome systems and research collaboration.

dr@sivitil.li

Based in Armagh, Northern Ireland. Open to relocation or remote work. EU, UK and US work eligibility.