I will make computer vision projects with opencv, mediapipe

I build AI products that move from demo to deployment.

HERE'S WHAT I CAN HELP WITH

Computer Vision

• Real-time video/camera AI; detection, segmentation & classification
• Pose/keypoints, OCR, inspection, anomaly detection & preprocessing
• Stack: OpenCV, YOLO, MediaPipe, PyTorch, U-Net, SAM, PaddleOCR

Inference & Deployment

• Low-latency serving, profiling, scalable APIs, Docker and GPU/cloud deployment
• Stack: Triton, TensorRT, ONNX Runtime, FastAPI, CUDA, AWS

Generative AI

• Custom image generation/editing, LoRA fine-tuning & Stable Diffusion pipelines
• Stack: ComfyUI, SDXL, FLUX, ControlNet, IP-Adapter

Document AI & LLMs

• OCR, structured extraction, RAG assistants, compliance, reports & review automation
• Stack: OpenAI, Claude, LangChain, Hugging Face, vector search

Selected Results

• 24 FPS posture-analysis product using custom keypoints
• Cut visual inspection from ~40 sec to ~1.5 sec
• TensorRT optimized CV pipelines from 7 sec to 1.5 sec
• OCR + LLM compliance reviews cut from ~1 hour to <15 min
• Triton model-serving and scalable inference deliveries

Let's connect!