Software Developer
Hi, I'm Mateusz
I'm a Python
developer
Python developer and Computer Science student at AGH UST, building backend systems, AI tools and full-stack web applications. Working with FastAPI, Django, React and Next.js, with hands-on experience in computer vision, NLP pipelines and LLM-based workflows.
My Skills
I have been learning software development independently for almost two years, focusing mainly on Python backend development, web applications, data processing, and practical AI/ML projects. I enjoy building complete, working solutions - from API design and database integration to frontend interfaces and deployment basics.
Backend
- Python
- FastAPI
- Flask
Frontend
- React
- TypeScript
- JavaScript
Databases
- PostgreSQL
- MongoDB
- MySQL
Cloud & DevOps
- Docker
- AWS
- GitHub Actions
Tools
- Git
- Linux
- VS Code
- Postman
My Projects
Selected work spanning NLP assistants, computer vision, full-stack web platforms, and machine learning dashboards - from API design and model training to deployment and user-facing interfaces.
Machine learning web application for forecasting air quality in Kraków using historical and near real-time station data.
- Built FastAPI backend for collecting and serving air quality data from KaggleHub and Airly API
- Implemented data processing pipeline for PM10, PM2.5, NO2, O3, SO2 and CO measurements
- Designed PostgreSQL schema for storing station metadata and time-series measurements
- Developed Next.js frontend dashboard for visualizing forecasts and station-level data
- Containerized backend with Docker and Docker Compose for reproducible local deployment
- Python
- FastAPI
- Pandas
- Scikit-learn
- PostgreSQL
- Docker
- Next.js
- TypeScript
Python application enabling natural-language control of Google Calendar using Gemini models.
- Built NLP-driven assistant translating user prompts into structured calendar operations
- Designed prompt-parsing pipeline for extracting event attributes (title, duration, reminders, color)
- Integrated Google Calendar APIs with validation layer and REST-based workflow
- Implemented create/edit/delete event automation with structured input schema
- Developed modular architecture prepared for Gmail automation extension
- Python
- Google Cloud APIs
- Gemini API
- REST integrations
Computer vision pipeline for detecting Varroa destructor parasites on bees using YOLO-based object detection.
- Curated and labeled ~1k bee images from public datasets using Label Studio
- Trained YOLOv8 model achieving up to ~80% detection accuracy with confidence scoring
- Built FastAPI inference service for image upload, model execution, and structured output
- Implemented end-to-end workflow: dataset preparation - training - REST inference endpoint
- Python
- YOLO
- PyTorch
- FastAPI
- Label Studio
Full-stack web platform for trading graphic cards built with Django and Next.js (TypeScript).
- Designed JWT authentication with HttpOnly cookies, including secure login and registration
- Developed responsive customer interface with shopping cart, product listings, ratings, and pricing
- Built advanced product filtering (price, brand, availability) and autocomplete search
- Created protected REST API endpoints with input validation and role-based access logic
- Modeled relational MySQL schema with many-to-many relationships for products, users and orders
- Django
- Next.js
- TypeScript
- REST API
- JWT
- MySQL
Education
AGH University of Science and Technology
B.Eng. in Computer Science in Engineering (2023 - Expected 2027)
Coursework includes statistical data analysis, optimization methods, databases (SQL & NoSQL), software engineering, algorithms and data structures, and object-oriented programming (C++).
Academic Projects
FEM Simulation
Engineering-focused simulation solving a transient heat conduction problem with internal heat sources using the Finite Element Method.
Member of Analytica Student Club at AGH
Scientific Activity (2025 -)
Working on an NLP-based system for extracting geodetic parameters from user input to support land-plot search. Built prompt parsing and data validation workflows, and integrated multiple LLM models for conversational interaction, intent recognition, and structured data extraction.
Hackathons
- HackYeah 2025 - Participated in a four-person team at the HackYeah hackathon, where we developed a full-stack retirement simulation system for the Polish Social Insurance Institution (ZUS).
- Kościuszkon 2026 - AI and full-stack team project combining machine learning workflows with a working web application.
- HERE Technologies Hackathon 2026 - Team software development challenge focused on building a practical solution under hackathon time constraints.