At PicsArt Armenia, I work on production-ready ML solutions for image generation and editing, including 4K style transfer, outpainting, and object removal.
I train and optimize diffusion and flow-matching models, and have created a differentiable 3D renderer with advanced material properties such as reflectivity. My work also includes face shape detection and manipulation using warping techniques.
I use Python, PyTorch, and OpenCV for ML and image processing, and C++ for optimized or algorithmic code. I also taught ML and DL at Yerevan State University.
