Adam Cofała

GPU-accelerated signal processing library developed during an internship at the Space Research Centre of the Polish Academy of Sciences. It was used to study ionospheric ionization and its changes over time by synchronizing three USRP channels, processing Digital RF I/Q recordings with CUDA, and generating spectrograms, cross-spectra, coherence, phase, and amplitude data for further analysis.

Gravitational N-body simulation rendering 30,000 particles in real time. Galaxies collide, spiral, and merge under gravity. Uses octree spatial partitioning to keep physics calculations fast and OpenMP to distribute work across CPU cores. Interactive 3D camera, dual galaxy presets, and color modes to visualize velocity and mass differences.

An IoT device designed to monitor and maintain the perfect environment for sourdough starter. Features temperature control using a Peltier module, real-time dough volume tracking with ultrasonic sensors, and automated email notifications when the dough is ready.

Interactive 3D geometry visualization tool designed for people who struggle with spatial thinking. Provides rotation, scaling, and inspection of 3D shapes from every angle to build intuition. Built with Three.js and TypeScript, with live deployment to GitHub Pages so updates go live automatically. Helps bridge the gap between mathematics and visual understanding.

I gained insight into the field of Artificial Intelligence through hands-on implementations and practical experience. These papers and projects are the direct result of that work, covering topics like Genetic Algorithms, Gradient Descent, Minimax, Neural Networks, and ID3 arrays.

Mandelbrot set renderer implemented with a mix of C++ and RISC-V assembly, tested on real RISC-V hardware. During studies, I had the opportunity to run this directly on an HiFive P550 board from SiFive, providing hands-on experience with actual low-level hardware optimization. Demonstrates assembly language performance techniques for mathematical computations with CMake cross-platform build configuration.

A pixel-art isometric colony survival simulator set on Mars. Build a self-sustaining settlement, manage energy and water logistics, expand pipe networks, and explore procedurally generated terrain with reproducible seed-based worlds. Current alpha features include binary save/load support, slope-aware rendering, and resource simulation across the colony.

Newton fractal renderer implemented in C++ with an optimized x86-64 (NASM) inner loop. Interactive OpenGL viewer with ImGui controls for zooming, iteration limits and real-time exploration.

Desktop app helping nature photographers plan shoots by combining weather forecasts with water level data. Includes 48-hour weather predictions and real-time river conditions for saved locations across Poland. Built on Folium maps for location discovery, Matplotlib for charts, and integrates the Polish Hydro Data project to pull live measurements from 500+ monitoring stations.

Flocking simulation with 10,000+ boids following Craig Reynolds' rules—they separate to avoid crowding, align with neighbors, and move toward their center. Emergent flock behavior appears from simple steering rules. GPU instancing renders thousands of agents smoothly, spatial grid optimization keeps interactions fast, and ImGui controls allow real-time tuning of behavior weights.

Conway's Game of Life showing how simple rules create complex patterns. Classic mode with familiar still-lifes and gliders, plus extended "Larger than Life" variant that produces traveling waves and intricate structures. Native Windows rendering with OpenMP parallelism for smooth performance on large grids. Includes pattern library (Gosper Glider Gun, rings) and pattern generation.

Simulation of 1,500+ bouncing spheres colliding elastically with proper physics. Spatial grid partitioning keeps collision detection efficient even with thousands of bodies. Built with NumPy for fast physics calculations and PyOpenGL for GPU rendering. ImGui control panel provides gravity and collision parameter adjustments and velocity-based coloring in real time.

Automated data pipeline pulling real-time water levels from 500+ Polish monitoring stations (IMGW API) and storing them on GitHub. Runs on schedule via GitHub Actions—no servers, no costs. Data updates automatically multiple times per hour and stays available as JSON for anyone to use. Used by my Photography Weather App to show photographers current water conditions at shoot locations.