Projects Overview
I’ve been involved in a variety of projects, ranging from hackathons to personal or university projects. Sometimes the goal was to try out new technologies, challenge myself, simply have fun or all together.
Machine Learning
Path Searching algorithms
This project compares pathfinding algorithms for public transport routes (MPK) using Python. It focuses on using Dijkstra and A* to find the best connection between two stations and explores the Tabu algorithm for the optimal route through multiple stations.
Minimax
This project compares various implementations of the minimax algorithm in GoLang, including recursive, concurrent, and alpha-beta pruning versions. It also analyzes different heuristics and parameters. The algorithm is demonstrated using Halma, a Chinese checkers variant where the goal is to reach the opponent’s base first.
EEG classifier
This project involves a pipeline that processes EEG experiment data. It lazily loads the data, preprocesses it, splits it into 1-second intervals, shuffles the data, and then uses it to create mental state classifiers with PyTorch.
Genetic Algorithms
Implementation of genetic algorithm that finds optimal solution for a knapsack problem in C++.
Brain waves analysis
This project uses R to analyze EEG data from “a study on mental state classification”, covering relaxed, neutral, and concentrating states. It includes artifact removal (bandpass and notch filters), feature selection with PSD, and baseline signal extraction.
data analysis classes
Web & Mobile
Interactive Map of Bieruń
Following Citified win at Hack The Climate 2023, my team delivered the interactive map of Bieruń within 7 months. As the lead software engineer, I managed the entire technical process: designing the backend in Nest.js, DevOps tasks like configuring nginx and orchestrating containerized services, and frontend React.js implementation including map functionalities, SEO enhancements and prerendering with Node.js. Additionally, I handled client communication and requirements gathering.
WorCare
I contributed to developing WorCare, a healthcare app for employers. My role focused on building the Next.js frontend and backend with MongoDB. Additionally, I uniquely pitched our idea by incorporating Jan Kochanowski’s “To Health” trifle. Our innovation earned 2nd prize in the Best Hacks 2023.
VoiceMess
A mobile app designed to read SMS messages aloud using a customized voice generated by ElevenLabs’ cutting-edge cloning service. Developed with Expo (React Native), it utilizes my Expo’s SMS module, crafted specifically for this project, to listen to incoming SMS (Android-only). User authentication is integrated through Facebook SSO.
MackawAI
A React-Native app designed to help users learn to speak foreign languages, powered by Google TTS, Speech-to-Text, and GPT-3, even before it became trendy. For the backend, we utilized Firebase.
Expo SMS
A Kotlin module for sending and receiving SMS messages with bindings in JavaScript, designed to work seamlessly as an Expo module.
Android projects
A collection of Android projects where I explored various Android development concepts, including Fragments, Activities, Layouts, Navigation, Drawers, and the Jetpack Compose Toolkit.
History Compass
We developed History Compass, a React-Native app that uses an API to provide information about monuments. The app shows the nearest monuments and redirects users to detailed pages for more information, comments, or saving for later. The challenge was addressing the lack of comprehensive records, fading memories, and the tendency to associate sightseeing only with travel. Our solution encourages exploration, engages the community, and brings information about Polish heritage closer to the public.
Questapp
React.js platform with Flask backend.
C#
Bore No More
Pipelon the biggest project, and also our bachelor thesis, that focuses on creating a C# framework for seamless BCI integration, that facilitates device connection, EEG signal classification, callbacks invocation and Unity integration.
Tetris BCI Experiment
Self-contained Unity experiment that enables easy EEG data collection and annotations. Created as part of the Bore No More project
TowerDefense
Unity PoC, enhanced by dialogues written with Ink language.
Others
EEG-collector
EEG-collector is a Python application designed to simplify the collection, visualization, and annotation of EEG data. Built with Tkinter and Matplotlib, this simple GUI facilitates EEG data collection while adhering to format standards and storing annotations in JSON.
NoteScribe
Note Scribe is an Python application for scribing notes from documents and articles using LLM technology. It leverages LangChain for document splitting, chaining and templating prompts, MenuAgent implementation, and semantic similarity searching. Embeddings and LLM (OpenAI) handle decision-making, summaries, QA, and note creation. Chroma is used for storing documents as embeddings, Tesseract for extracting text from images or unselectable PDFs, and Eleven Labs for converting notes into high-quality, natural-sounding podcasts.
Dummy-Generator
This project automates the tedious task of manually entering large amounts of data into MS Access. Using the Faker Python library, I created a data generator that streamlines this process, saving time and reducing the monotony for my university peers.
Auto-meetings
A tool that ensures you never miss meetings by managing your schedule and waking your PC before each meeting. Written in Python, HTML, CSS, Shell. It features a WebUI where you can view upcoming meetings and even check your GMail mailbox for new ones. The app runs as a lightweight systemd service, allowing you to hibernate your PC and have it wake up automatically to record meetings or check for updates. Installation requires basic terminal commands, and the app is free and open-source.