Skip to content
View markolivaic's full-sized avatar

Highlights

  • Pro

Block or report markolivaic

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
markolivaic/README.md

Earth's land as points, with 18,983 tracked objects moving in their orbital planes

Software developer in Zagreb and Linz. I build systems, then measure whether they work.

now   Ericsson Nikola Tesla   Java for the framework that validates the 4G/5G radio network
      Vectra XR               co-founder and CTO, 3D configurator for fitness racks and prefab homes
      JKU Linz                MSc Artificial Intelligence, 2026 to 2028
project what it does measured
surrey lidar point-cloud segmentation, PointNet++ style 0.93 point accuracy on held-out tiles, 3 seeds
kessler satellite close-approach screening with SGP4 18,983 objects over 24 h, every pair within 25 km
asymmetric-society LLM agents in iterated economic games 320 games, pre-registered on OSF
anneal live CHGNet crystal relaxation in the browser every relaxation step streamed, CPU only
suncokret roof-by-roof shadow study of central Zagreb from the city's own LoD 2.2 3D model

livaic.vercel.app | LinkedIn

Pinned Loading

  1. asymmetric-society asymmetric-society Public

    Capability-asymmetric LLM agents in iterated economic games: who exploits whom, and does asking agents to commit in public flip the capability premium? Pre-registered (OSF).

    Python 1

  2. surrey surrey Public

    Trains a PointNet++ style lidar segmentation model on Surrey, BC (DALES-2), runs it on USGS 3DEP San Francisco, and scores each area against footprints, NAIP vegetation and nulls.

    Python 1

  3. suncokret suncokret Public

    Computes rooftop solar potential across a Zagreb district from the city's LoD 2.2 building model and PVGIS irradiance, including shading from neighbouring buildings.

    Python 1

  4. kessler kessler Public

    Screens 18,983 catalogued objects with SGP4 for close approaches over 24 hours, filters two classes of false conjunction, and re-propagates a satellite after a delta-v burn.

    Python 1

  5. sada sada Public

    One tap. Real shadow physics. Sada tells you the best spot to sit in Zagreb right now. Sun or shade, cafe or bench.

    TypeScript 1

  6. anneal anneal Public

    Loads Crystallography Open Database structures, filters them for CHGNet, and streams live relaxations after substitution, vacancy or strain edits. 27.8% of COD qualifies.

    Python 1