LinkedIn: www.linkedin.com/in/hubert-jastrzebski

I'm a student at Jagiellonian University, studying Theoretical Computer Science

• 🖥️ Passionate about software engineering, machine learning, and AI research
• 🦾 Skilled in algorithms, data structures, and problem-solving
• 🥋 Training jiu-jitsu since the age of seven
• 📖 Fan of Brandon Sanderson’s books
• 🧶 Enjoy crocheting in my downtime

Developed a custom chess engine featuring two distinct bots: an AlphaZero-style reinforcement learning agent and a supervised learning model trained to predict human moves. The supervised approach achieved over 57% move prediction accuracy and an estimated Elo rating of 1,900.

Co-authored a research paper on watermarking Latent Diffusion Models (LDMs), accepted at ICLR 2026 (pdf). Engineered the core PyTorch codebase and orchestrated the execution of large-scale experiments on remote GPU clusters.

A desktop application for designing and training simple neural networks. Built from scratch in Java using a clean MVVM architecture, featuring a custom neural network engine, an intuitive JavaFX interface, and a PostgreSQL database for managing users, models, and training data.

Ongoing research project analyzing dependencies in mother-child facial expressions during interactions. Utilizes OpenFace2 to extract facial Action Units for complex temporal and relational analysis.

Co-authored a paper on continual learning, accepted at Knowledge-Based Systems (pdf). Designed and implemented the core algorithmic logic, generated data visualizations, and wrote the results section in the paper. FeNeC is a continual learning method that employs feature clustering and neighborhood-based classification with frozen feature extractors to mitigate catastrophic forgetting.

A lightweight C++ library inspired by PyTorch for building basic deep learning models. Implemented core automatic differentiation, backpropagation, and gradient descent optimization algorithms completely from scratch to explore low-level machine learning mechanics.

An Android life-simulator game developed in primary school, previously published on Google Play. Players navigate life choices, such as pursuing a career or building relationships, while managing basic resource needs.

A comprehensive archive of my algorithmic journey and competitive programming solutions. It features implementations of advanced data structures and algorithms, including solutions to over 700 tasks from Olympiad in Informatics preparations and university coursework.