I'm Tcherno

Fully open-source, end-to-end Data Lakehouse platform designed to unify high-performance storage, real-time stream processing, semantic transformation, and intelligent orchestration

Multimodal Retrieval-Augmented Generation (RAG) system for complex financial report analysis using Claude Vision, BGE-M3, OpenSearch hybrid search · 84.4% RAGAS score

A Django web application for managing and visualizing data table metadata and lineage relationships, backed by a Neo4j graph database

Visual question answering system using the ViLT model that allows users to analyze images and automatically answer questions about their content

Neural style transfer web application using the VGG19 model to apply artistic styles to images and create visually stunning results

Django web app that classifies text as AI-generated or human-written · Logistic Regression + TF-IDF + spaCy feature engineering · trained on 260K samples

Developed a web application that combines signal processing techniques with conversational AI to analyze and interpret audio data

Complete pipeline with preprocessing, modeling, tracking (MLFlow), monitoring and deployment - Performed customer segmentation analysis on the Olist e-commerce dataset using unsupervised learning techniques to identify distinct customer groups and inform targeted marketing strategies

Developed a neural style transfer web application using the CycleGAN model to apply artistic styles to images