I design and develop end-to-end generative AI systems — from model fine-tuning to integration in real-world applications. My work focuses on building stable, modular pipelines that produce consistent, high-quality results at scale. I work with node-based workflows, combine multiple models within unified systems, and optimize generation processes for speed, control, and reproducibility. My approach goes beyond prompt-based generation, focusing on engineered solutions for production environments.
With a background in VFX, AR, and digital production, I bridge the gap between creative and technical implementation. My expertise spans the full AI lifecycle — from model fine-tuning to automation and deployment in user-facing applications. I thrive on exploring new technologies and turning cutting-edge AI research into practical, production-ready solutions.




Development and training of custom LoRA models for stylistic, character, and contextual adaptation. My workflow includes dataset curation, tagging strategies, training optimization, and seamless integration into generative pipelines (ComfyUI, Flux, SDXL).
Architecture and development of advanced graph-based generative systems in ComfyUI. I specialize in constructing modular, scalable node pipelines that combine multiple AI models and tools, enabling automated, high-performance image and video generation workflows