Filter skills (0)
Skills & Technologies
Dávid József Schofhauzer
Professional Experience, Education & Projects in Software Development
Open to work.
Software developer with experience in full-stack web development, AI systems, and infrastructure. Passionate about building reliable, maintainable, and scalable software solutions. Skilled in Node.js, Python, React, and cloud technologies. Strong problem-solving abilities and a collaborative mindset.
Work Experience
Sonepar Hungary Kft. - Veszprém, Hungary
Comparison, prototyping, documentation, and presentation of OCR solutions.
Conducted a technical comparison of OCR solutions, evaluating accuracy, performance, and integrability. Developed prototypes to test the solutions and documented the results. Presented findings to the development team to inform decision-making.
Education
Bachelor's degree in Computer Science
Completed a Bachelor's degree in Computer Science at the University of Pannonia, gaining a strong foundation in programming languages such as SQL, C++, C#, Java, Javascript, Typescript, HTML, CSS, Python and Blazor. Acquired knowledge in database administration with MySQL, operating systems including RedHat Enterprise Linux, computer networks, network security, artificial intelligence and machine learning. Developed practical skills through coursework and projects, preparing for a career in software development and IT.
Highlighted Project Case Studies
These projects represent my strongest full-stack, AI-system, and infrastructure work. Use the skill filters on the left to narrow what you want to evaluate.
Built a public, open-source MCP server that exposes live Helldivers 2 game data to LLMs, deployed as a stateless edge service on Cloudflare Workers.
Built a full-featured community website for a Helldivers 2 themed Discord server using React 19 and TanStack Start, deployed as a server-side rendered Cloudflare Workers app.
Built an AI service that routes requests across Claude and DeepSeek and coordinates tool calls through MCP.
Implemented a tool registry and execution service that exposes model-callable functions for database queries, web queries, and integrations.
Developed a production-like Discord bot platform with slash commands, deferred interactions, gateway event handling, and internal REST endpoints.
Built an Express and PostgreSQL backend with service-layer architecture.
Designed and operated a self-hosted, Docker Compose based stack with reverse proxying, secure ingress, CI/CD automation, and service monitoring.
Additional Projects
Unity-based fitness game integrating real-world activity via Google Health Connect.
Created a Unity fitness game that uses Google Health Connect data to influence gameplay. Built a custom Unity plugin that provides a C# to Kotlin bridge for Android integration.
University team project for dormitory issue reporting with a full-stack web architecture.
Built a bug reporting web application for dormitory residents in a 4-person university team. Worked on backend modelling, database design, REST API implementation, and frontend development.
Personal infrastructure project running production-like services on a repurposed Ubuntu server.
Repurposed an old PC into a Ubuntu home server hosting NAS and web services behind Nginx reverse proxy with Cloudflare Tunnel access. Set up Docker and Docker Compose workloads, automated cron backups, and monitoring. Live services include Node.js Discord chatbots, an AI inference service, an MCP service, a PostgreSQL and Redis backend for a Discord community server, a GPU-accelerated image processing pipeline, and a GitHub webhook-driven CI/CD pipeline.
University team project for hand gesture classification and gesture-driven mouse control prototyping.
Developed a hand gesture recognition solution using a CNN-based classifier in a 3-person team. Implemented a prototype for gesture-driven mouse control and integrated the recognition workflow into the application pipeline.
Personal project training and tuning a small GPT-2 style model on a custom dataset.
Trained a small GPT-2 architecture with approximately 1.8 million parameters on a custom dataset. Documented the training process, tuned hyperparameters, and configured inference runs.