Mykhailo Medved

Built a large-scale Product Retrieval System for fashion e-commerce, achieving 90% accuracy in top-10 results across thousands of products. Developed an E-commerce recommendations pipeline with retraining, monitoring, and alerting. Created an OCR and Document Processing Agent using LLMs and OCR for automated invoice and specification extraction, handling large volumes of documents daily.

Developed web/mobile app for Generative AI content creation. Built AI microservices for text and image generation, managed GPU workloads, fine-tuned generative models, and integrated multimodal LLMs. Implemented a Retrieval-Augmented Generation (RAG) chatbot over 10,000+ PDF specs, with reindexing and embedding pipelines, integrated into an e-commerce platform.

ERP B2B Platform: unified platform for payments aggregation and integrations, including ML solution for sentiment analysis and 3DS payment integration. Image Classification Module: developed an object detection and classification system using YOLO with synthetic data augmentation, integrated into a larger product retrieval system.