Roman Meszaros

An AI agent that learned a strategy of a board game. I designed the game environment for the agent to learn in and a rule-based agent that the AI learned from. The AI agent used feedforward neural networks and reinforcement learning, and I trained the AI agent against different rule-based agents. I designed this as part of my dissertation project.

A full-stack addition to an existing Drupal module. The project started with the client's initial idea and research. The goal was to support researchers when working with ancient manuscripts, enabling them to order different manuscript readings according to their reliability.

I designed the app to guide researchers when ordering manuscripts. The ordering happened in two steps: bucket sort and merge sort. In each step, the researchers made the atomic decisions to put an item into a category or compare two items as the app was prompting them to make the relevant decisions. I designed the algorithms and database structures and created RESTful API services and the React front-end app.

A personal study project, building a full-stack multiplayer board game website. The game front end is created using React. The back end was done in PHP using the CakePHP framework. The back end exposed restful API services to which the front end connected. The app was based on an actual board game.

The Toptal React Accelerator is an exclusive learning program that educates JavaScript developers in React, NextJS, Redux, and GraphQL. This is a 2-month program that consists of rigorous learning exercises and hands-on activities, a stringent evaluation process, comprehensive learning assessments, and 30-40 hours of building an Invoice Management System. Users can track invoices, clients, and company information within this invoicing application.