Bohdan Horvat

Developed an ML-powered automated product matching system for e-commerce with 1M+ product database, achieving 95%+ accuracy with <1 second response time. Built end-to-end data pipeline on Google Cloud Platform (Vertex AI, Vector Search, BigQuery), created FastAPI service with semantic search capabilities, and set up production-ready environment with Docker containerization and Grafana monitoring. Implemented comprehensive solution from data processing to vector search using cosine similarity.

Automated data collection and processing workflows by migrating complex analytical system from Google Sheets to professional ETL platform. Built data aggregation system from competitors, external APIs and internal sources with subsequent normalization and business logic implementation. Configured Airflow scheduler for regular data updates with flexible schedules including monitoring and alerting.

Developed data collection and processing solutions at consulting company, working with multiple clients. Created and refactored web scrapers using Python (Scrapy, requests, BeautifulSoup), developed API services for data access. Performed data normalization and structuring according to client requirements, conducted data analysis. Maintained microservices architecture and ensured stable operation of parsing systems.