Alexei Czornyj

I’m a Master’s student in Software Design and Development at Université de Poitiers.

My work spans neural importance sampling for layered materials, GPU-accelerated interactive visualisation, and physically-based rendering.

Projects

Kestrel

A physically-based CPU path tracer written from scratch in C++. Reads Mitsuba-style XML scenes and renders them with unidirectional path tracing, Multiple Importance Sampling (MIS) and Next Event Estimation, in a modular architecture. Documented with MkDocs and Doxygen.

C++ Path Tracing MIS Rendering

Docs GitHub

Material Explorer

An interactive analytical BSDF viewer built in C and Vulkan, developed at the XLIM Laboratory. Enables real-time GPU-accelerated visualisation and manipulation of analytical material models, with efficient parallel computation for interactive performance. I contributed to this project as part of my research internship.

C Vulkan BSDF GPU

Open source soon

MyLittlePlanet

A self-contained procedural planet renderer in WebGL. Adapted from my computer graphics coursework.

WebGL GLSL FBM Procedural

Live GitHub

CV

Education

Master EUR Software Design and Development — Université de Poitiers (Sept 2024 – Sept 2026)

Geometric design & simulation for health, machine learning, advanced 3D algorithms, geometric data structures, parallel programming.

Licence Informatique (BSc Computer Science) — Université de Poitiers (Sept 2021 – June 2024)

Object-oriented & functional programming, software engineering, human-computer interaction.

Research Experience

Research Intern — École de Technologie Supérieure, Montréal (April 2026 – Aug 2026)

Extension of neural importance sampling for multi-layered materials; study of generalization across material classes and robustness in Monte Carlo integration; Mitsuba integration with custom BSDF and sampling plugins. Advisors: Adrien Gruson, Mickaël Ribardière.

Part-time Research Intern — XLIM Laboratory, Poitiers (Sept 2025 – March 2026)

Interactive C/Vulkan visualiser for material data with efficient GPU parallel computation; real-time manipulation of analytical models. Advisor: Mickaël Ribardière.

Research Intern — École de Technologie Supérieure, Montréal (June 2025 – Aug 2025)

State of the art implementation of neural networks for multilayer material importance sampling; custom BSDF integration in Mitsuba. Advisor: Adrien Gruson.

Part-time Research Intern — XLIM Laboratory - ANR Digitalis (Sept 2024 – May 2025)

Texture synthesis algorithms in Python; artwork ageing simulation via example-based methods. Advisors: Rémi Orveau, Daniel Meneveaux.

Research Intern — XLIM Laboratory - ANR Digitalis (April 2024 – June 2024)

Analysis and visualisation of data from point-cloud rendering simulations; large-scale data processing with NumPy and Pandas. Advisor: Mickaël Ribardière.