Building intelligent systems that actually work 🚀
I build AI-powered products — from computer vision pipelines and demand forecasting microservices, to full-stack mobile apps with LLM integration. Based in South Korea 🇰🇷, building for Uzbekistan 🇺🇿 and beyond.
AI demand forecasting microservice for Uzbekistan ERP market. Custom ML engine built from scratch with pure NumPy — no sklearn. Predicts 30/60/90-day revenue with confidence bands, Uzbekistan holiday calendar (Navro'z, Ramazon). Full Java Spring Boot integration. Interactive Chart.js demo dashboard.
Full-stack mobile app where users generate original music by describing a mood, GPT analyzes lyrics to detect emotion, and matching dance choreography is recommended. 2025 Capstone Design project. React Native + Spring Boot + AWS S3 + Community feed.
Two-stage cascade pipeline for real-time bottle/can detection using Oriented Bounding Boxes — tighter, rotation-aware boxes. Balanced dataset re-split, FastAPI REST API with /predict endpoints. Tested across 6 real scenes: outdoor, café, in-car. Trained on Colab T4 GPU.
YOLO11 + ByteTrack vehicle tracking with two-line speed measurement. Blue line = entry timer start, Yellow line = exit, speed = distance/time. Color-coded: green under 40, blue 40-80, red over 80 km/h. iPhone Continuity Camera support. No training required.
Real-time person detection via MacBook/iPhone camera using YOLO11n — the fastest and lightest variant. Multi-language README (UZ/KR/EN).
AI Python Full-Stack course project from Mashaqqat bootcamp. Backend + AI integration practice.
Open to collaborations, AI consulting, and interesting projects. Feel free to reach out!