Szallós Kis Csaba (1998) is a software developer with an appetite for natural language processing, machine learning and data visualization.

He also writes poems in Hungarian.

Bio

studied computer science as BSc at the Babeș-Bolyai University between 2017 and 2020.

studied data analysis and modelling as MSc at the Babeș-Bolyai University between 2020 and 2022.

worked as a C++ software developer from 2019 until 2024.

working as a Python software engineer since 2024.

Projects

arcplot - a published Python library based on matplotlib for seamless creation of custom arc diagrams.

visual haikus - a Python application which writes haikus based on images. the goal of my license thesis: creating a program which writes a poem in the form of haiku from a landscape, a painting, a sketch, or any visual input. thus the program would 'behave' as a poet, it would create art from a source of inspiration. built ML model and attempted to find a possible generative grammar for haikus. technically speaking: a multi-label image recognition in Keras combined with dependency parsing of different texts with spaCy.

onrepeatify - a Python application for automatically adding songs from automaticvally generated 'On Repeat' Spotify playlist into a 'favorites' playlist collecting the all time favorites.

interactive map of extinct and endangered languages - a webpage which displays extinct and endangered languages all around the world. based on the Leaflet framework the goal was to create a data visualization which can help users navigate easily while exploring languages. by clicking on their names, the appropriate wikipedia article will pop up for a more in-depth understanding.

spellcheckmate - a hopefully cross-platform reimplementation of Peter Norvig's spelling corrector with minor modifications.

geometric tunes - interactive new media installation written in python which was showcased at Clujotronic electro arts festival deployed on a Raspberry Pi.

interactive visualization of the best movies based on letterboxd - using Flourish to represent the top 250 narrative feature films based on letterboxd.

rhythm - metric foot transcription applicable to Hungarian poetry initially written in java and afterwards in javascript.

Braille OCR - a cross-platform application which can identify Braille-characters written by a Braille-typewriter. in order to assist the educational process for the visually impaired, the aim of my dissertation thesis is to develop a smartphone application which monitors a mechanical Braille typewriter and recognizes the Braille-characters in real time. my main focus: creating the ML model which performs OCR on a video recorded by a smartphone in real-time.

ï