Projects
ReelMate
Developed an AI-driven platform that streamlines the creation and distribution of short-form video content across major social media platforms. Engineered features to transform text-based content into engaging videos, automate cross-platform scheduling, and provide analytics to optimize engagement. Integrated a Telegram bot to facilitate seamless content approval workflows. Utilized FastAPI and AsyncIO to ensure high-performance, real-time operations.
AIM Hackathon - TIMETOACT Group 2nd Place Winner
A platform with Retrieval-Augmented Generation (RAG) to identify greenwashing in ESG reports.
susteam - Full-stack development of ESG platform
Architecture and development of the ESG platform for generating sustainability reports in the hospitality industry.
Rail4Future Data Science Project 2024
Analyzing rail irregularities for industrial partners Siemens and ÖBB.
Full-stack Development of the NFT Platform
Implemented a secure authentication system using MetaMask, which allows users to store and manage seed phrases as encoded NFTs. Developed the platform’s backend to interact seamlessly with the Polygon blockchain, ensuring efficient and secure transactions.
LawSyncr
Led the end-to-end full-stack development of the project, including the implementation of a Minimum Viable Product (MVP) with integrated machine learning components.
Crypto Trading Analysis and Future Trading
Engineered a bot to systematically scrape and collect trade data from Binance, allowing detailed real-time market analysis. Automated the execution of trading strategies using a custom-built bot, ensuring consistent and timely trades.
Portfolio Website Development
Designed and developed a portfolio website, integrating continuous integration/continuous deployment (CI/CD) pipelines for seamless updates via GitHub Pages. Utilized Obsidian vault data to create an interactive front-end that visualizes my professional journey using graphology and sigma libraries.
Document Analysis and Optical Character Recognition (OCR)
Conducted a comprehensive layout analysis of historical documents, identifying key structural elements to improve text extraction accuracy. Integrated advanced Named Entity Recognition (NER) using SpaCy and fine-tuned language models (e.g., LLaMA-3) to extract metadata from images.
Lip Reading app for Austrian Red Cross
Developed a lip reading app using AV-HuBERT for the Austrian Red Cross to communicate in noisy environments. Integrated the model into a web app, passing the video stream from the smartphone to the model, reading the lips, and generating text.
Data Pipeline and Knowledge Graph for Tourism
Implemented a data pipeline mapping data from different sources and building a knowledge graph for tourism subjects. Visualized the results as an interactive map with a chatbot to analyze points of interest.
GLOSAV - Machine learning analyst
Engineered a sophisticated algorithm to detect railway routes with object detection and tracking mechanisms. Developed a robust way to track obstacles on the rails.
Printerra
Diploma thesis ”3D printer failure detection system based on computer vision detection”.