Demetrius
AI engineer specializing in large language models, image generation, and machine learning automation. Building the future of intelligent systems with cutting-edge technology and innovative solutions.
View My WorkAbout Me
I'm a passionate AI engineer with expertise in large language models, computer vision, and machine learning automation. I specialize in building intelligent systems that push the boundaries of what's possible with artificial intelligence.
My work spans from training custom LLMs and developing image generation pipelines to creating automated ML workflows that scale. I'm deeply fascinated by the intersection of research and practical applications, turning cutting-edge AI research into production-ready solutions.
When I'm not building AI systems, you'll find me exploring the latest developments in generative AI, contributing to open source ML projects, or experimenting with novel approaches to machine learning problems.
Featured Projects
AI Interior Design Generator
Automated script pipeline that generates photorealistic interior design and architecture material samples. Creates thousands of room shots with different styles, materials, and lighting conditions using advanced diffusion models and ControlNet guidance.
Artist Style Pipeline
Scalable automation framework generating millions of images across hundreds of artistic styles. Features intelligent prompt engineering, style consistency validation, and distributed processing for massive dataset creation with thematic coherence.
Real-Time Avatar System
Explorative research into real-time facial recognition and avatar generation. Implements advanced facial landmark detection, expression mapping, and live avatar synthesis with sub-100ms latency for immersive virtual experiences.
AI Mathematics Tutor
Advanced mathematics tutoring system combining LLMs with symbolic computation engines. Features step-by-step problem solving, validation against mathematical APIs, and adaptive learning paths for personalized education experiences.
Image-to-Video Pipeline
Cutting-edge pipeline converting static images to dynamic video renderings using customized video diffusion models. Implements checkpoint optimization, LoRA fine-tuning, and temporal consistency for cinematic quality outputs.
CV Rock Paper Scissors
Explorative study combining computer vision and machine learning for real-time rock paper scissors recognition. Features custom hand gesture dataset collection, CNN training pipeline, and live webcam inference with accuracy optimization.