Giovanni Cozzolongo

Embedded software | Satellite systems

C++ · Linux · Python · MATLAB/Simulink

At VersoSat, I develop C++ device-control software for Linux and satellite communications models in MATLAB/Simulink. At Tumbleweed Mars, I build a Unity/C# rover simulator and Python navigation tools to reduce localisation error without GPS.

Based in Erlangen, Germany. Open to roles across Europe. Available to start as soon as possible.

Email me CV (PDF)LinkedIn

Selected work

Embedded softwareVersoSat

Device control and communications modelling

Enabled device-control testing before board integration and took a DVB-S2 transmitter model through HDL generation and synthesis.

Built a C++ library for controlling radio hardware from Linux applications and adapted MATLAB/Simulink transmitter models for integration. VersoSat exhibited at IAC 2026 in Antalya.

C++ · Linux · libmetal · AArch64 · MATLAB/Simulink · Fixed-point modelling

Read case study

Simulation softwareTumbleweed Mars

Simulating a wind-driven rover in Unity

A configurable Unity/C# testbed with virtual sensors and reference trajectories.

Developed a Unity simulator in C# to study rover motion and generate data for localisation development. The navigation work features in the IAC 2026 interactive programme. A talk at iSpaRo 2026 is scheduled for 4 November 2026.

C# · Unity · Physics simulation · Virtual sensors · Python · Linux

Read case study

More software work

Experience

VersoSat

Feb 2026 to present

Embedded Software Engineer

Italy · Remote · Part-time

Developed a C++ device-control library using the team's hardware-access wrapper, adding input validation, state checks and event callbacks.
Tested driver logic with simulated registers, verified Linux Userspace I/O access and cross-compiled the library for AArch64.
Adapted MathWorks DVB-S2 transmitter models and checked receiver integration, HDL generation and block-level synthesis.

Device control and modelling case study

Tumbleweed Mars

Nov 2021 to present

Robotics Software Engineer

Netherlands · Remote · Part-time

Developed PESTO in Python using encoder-aided sensor fusion and evaluated visual odometry separately. Built a Unity simulator in C# with rover dynamics, terrain and wind inputs, virtual sensors and reference trajectories for navigation development.
Refined relay architecture descriptions and diagrams around link capacity, onboard storage and satellite contact windows. The relay study is also included in the IAC 2026 technical programme.
Adapted a terrain-segmentation transformer for consistent classification across camera orientations. Mean intersection over union rose from 0.704 to 0.776 on 526 AI4Mars test images evaluated at eight angles.

Localisation case study Unity rover simulator

FAU Erlangen-Nürnberg

Oct 2023 to present

Research Software Engineer

Erlangen, Germany · Contract

Built C++ readers to export detector data for Python analysis, with automatic file closure, event-identity checks and output-size limits.
Built a PyTorch pipeline to search 3.1 million detector images for muon-ring candidates outside the standard calibration selection, with traceable predictions and separate training, validation and inference.
Applied clustering and model comparison to satellite data, shortlisting eight regions from 48 candidate groups. The catalogue analysis was published in Astronomy & Astrophysics (2026).
Built Python sky-map simulations and source-detection analysis to estimate which pulsar environments gamma-ray observatories could detect. The simulation study was published in the ICRC 2025 proceedings.

Catalogue analysis case study

Italian National Institute for Nuclear Physics (INFN)

Sep 2022 to Aug 2023

Data Scientist

Pisa, Italy · Internship

Built a Python pipeline to analyse 14 years of satellite measurements around seven recorded events.
Combined statistical tests and background-model checks to produce confidence bounds for testing signal-variability models. The signal-variability study was published in Astronomy & Astrophysics (2025).

University of Pisa

May 2021 to Jun 2023

Instrumentation Project Lead

Pisa, Italy

Proposed the project, secured funding and led the team that designed and built a solar tracker and spectrometer for measurements aboard a stratospheric balloon.

Technical skills

Embedded software: C++ · Linux · Register interfaces · Unit tests
Communications: MATLAB · Simulink · DVB-S2 models · Fixed-point interfaces
Robotics: Python · C# · Unity · Sensor fusion
Machine learning: PyTorch · Transformers · Computer vision · Model validation
Data engineering: Python · Data pipelines · SQL · Regression testing

Open source

Open Climate Fix

Machine Learning Contributor

Enabled MetNet-3 attention layers to use forecast lead time, with layer and full-model regression tests.

Merged contribution

Catalyst Cooperative

Data Engineering Contributor

Added energy-storage tables and validation checks to Public Utility Data Liberation (PUDL). A separate contribution extended extraction across 15 annual workbooks.

Storage outputs Workbook extraction