Technology has always fascinated me, especially its impact on society. This passion led me to study computer science, where I developed strong skills in programming and systems administration. I am organized, eager to learn, and focused on finding effective solutions. Currently, I am particularly passionate about machine learning and AI.
π’ Voice-To-Voiceless
Assistive communication platform for people who cannot communicate through speech, built by a team of 4. It combines an accessible visual communication board with gaze-based interaction, facial-expression input, and a patient-to-care-team notification channel.
Tech stack: MediaPipe, WebGazer, REST, WebSockets
My role: Full-stack engineer
- Web-based gaze estimation using the device camera, MediaPipe Face Landmarker, and WebGazer
- Nine-point calibration, head-movement compensation, temporal filtering, and dwell selection
- Patient linking and persistent notifications between patients and nurses over REST and WebSocket APIs
π Forma
Open-source desktop app for exploring, profiling, cleaning, and transforming large datasets through a visual interface. It helps data scientists and ML engineers prepare datasets without rewriting the same preprocessing scripts. All processing runs locally, with no cloud hosting, Docker, or external backend required.
Tech stack: React, TypeScript, Tauri, Python
- Cross-platform desktop application with a visual interface for dataset exploration, profiling, cleaning, and transformation
- React and TypeScript frontend, Tauri shell, and a Python data-processing layer
- Fully local processing with no external dependencies
π Bookwise
Full-stack book recommendation application. A React frontend sends natural-language questions to a FastAPI backend, which combines input filtering, semantic retrieval from a Chroma catalogue, OpenAI recommendations, local summaries, and Google Books fallback search.
Tech stack: React, FastAPI, Chroma, OpenAI, Google Books API
- Recommends books from the local catalogue using natural-language questions
- Refines recommendations by mood, genre, theme, or reading preferences
- Searches and prioritizes books in the language of the prompt or a language explicitly specified by the user
- Retrieves complete summaries for books in the local catalogue
- Falls back to Google Books search when the local catalogue lacks enough relevant results
- Continues conversations with bounded message history
- I've started learnign PCB design in my spare time
- I'm curently looking into integrating AI and Robotics π€
- My favourite game is Hollowknight
- Hobies: gym, basketball, gaming

