portfolio

We explored two fundamentally different approaches to digital clothing try-on. Our mission was to investigate both image manipulation and 3D mesh-based solutions, each serving distinct use cases in the virtual fashion industry.

Our mission was to transform raw architectural blueprints into realistic and accurate 3D representations that could be explored in any browser—or even with VR headsets.

Creating a seamless and secure environment where customers could browse, select products, and leave without traditional checkout processes.

We developed a universal medical assistant that stores up-to-date information from scientific medical sources and assists in clinical decision-making.

Custom protein structure classification, based on PDB and CATH data. The core task is predicting protein architecture based on its amino acid sequence and 3D structure.

This is a test assignment I did for a startup. I really liked these tasks and it looks spectacular - so I recommend taking a look

API code for detecting walls and rooms in architectural floor plans. I’ve converted CubiCasa5k dataset to COCO format for smoother integration with detection framework.

The module allows you to collect statistics and build a dashboard displaying screenshots of violations, as well as elementary graphs for the last 24 hours

This project fine-tunes a Gemma language model using PEFT and LoRA on a dataset of texts by communist authors of the Soviet era.