AMSYS
Technical Support Specialist
Supported Linux servers and Docker-based services, and used SQL to diagnose customer issues.
-
Linux
-
Docker
-
SQL
Hi! I'm Adam Juhás.
Python and Convex backends, a touch of DevOps,
and Next.js on the frontend.
I work at two startups, building software for digital pathology at AigoPath and accounting applications at MoneyQ. I'm studying in the master's degree programme in Software Engineering at FI MUNI.
WHAT I WORK WITH
AT WORK
Technical Support Specialist
Supported Linux servers and Docker-based services, and used SQL to diagnose customer issues.
Junior Software Developer
C# modules for a call center system, frontend and backend logic for a terminal application, and services on remote servers.
Software Developer
Python and FastAPI services for AI-powered digital pathology. I also design and deploy applications with Docker and Kubernetes.
Full Stack Developer
Accounting applications with Next.js and Convex. I use n8n to automate workflows and integrations.
MY PROJECTS
My bachelor's thesis at FI MUNI. A React interface and TypeScript server control an Arduino LED matrix through a real-time WebSocket connection.
A FI MUNI team project for drivers, races, cars and components. Four Spring Boot services communicate through REST and Artemis, with Docker Compose and Prometheus/Grafana monitoring.
I contributed to this academic planning app. It combines tasks, subjects, a timetable, project boards and study analytics in a Next.js application backed by PostgreSQL.
A four-person Java Swing recipe manager with H2 persistence, transactions and asynchronous XML import and export. Built for FI MUNI's PV168 course.
A seminar project for exploring films and booking seats. React and Tailwind on the frontend, with a TypeScript backend, PostgreSQL and Redis.
A card collection manager built with WPF and MVVM. It combines API data, filtering and sorting, with JSON storage for the collection.
MY TOOLBOX
AI IN MY WORKFLOW
I use Claude, Codex and Cursor every day. I keep testing new models against real development work, because each one is better at a different kind of task.
I follow model releases, run my own tests and compare notes with other developers. The goal is simple: use the model that handles the job best.
For larger tasks, I run coding agents in separate Git worktrees. While one writes code, I review another agent's changes and keep the work moving.
I maintain repository instructions and reusable skills so agents follow the codebase architecture, conventions and team guidelines.
Greptile and Codex review pull requests before merge. I have also added smaller, cost-efficient models to product automations where they are accurate enough for the job.
A LITTLE ABOUT ME
I watch for the racing, team strategy and engineering.
Long rides help me clear my head.
I build decks, collect rare cards and play with friends.
I lift weights. Simple as that.
Have a role in mind, a project to discuss,
or just want to talk tech? I'd love to hear from you.