Tung P.Private profile: Upwork shows its earnings only to signed-in members, so we check it every few weeks, not daily. MRR here is a rough estimate (Upwork's six-month earnings ÷ months active). Last read Nov 5, 2026.

You have images, documents or data, and you need an AI model that works on your real data, not just in a demo. I build AI systems end to end, from your raw data to a model running on a server or a device: computer vision, LLM and RAG, and data science. I'm a Kaggle Bronze medalist in medical image segmentation (top 5% on the world's leading data science competition platform), and I led the team behind an autonomous robot that ran 3 models at 40 FPS on a Jetson TX2 (national Third Prize). WHAT I CAN HELP WITH: 👁️ Computer vision: object detection, counting and tracking with YOLO; image segmentation, including medical scans; defect and damage inspection; running models on devices (quantization, TensorRT, NVIDIA Jetson) 🤖 LLM, generative AI and automation: RAG over your documents with sources, agent integration, n8n, Power Automate, fine-tuning vision-language models, image generation, AI Video Editing 📈 Data science: forecasting, predictive models, explainable ML, data visualization RESULTS FROM PAST WORK: - Improved the vehicle and road damage detector for an intelligent traffic monitoring system at Viettel, Vietnam's largest telecom group (+0.05 mAP50) - Segmented stomach and intestines on 38k+ MRI scans to support radiation therapy planning (Kaggle Bronze, top 5%) - Built an assistant that answers questions over 1,110 pages of EU reports and shows the source pages behind each answer (page search tested on 318 questions) - Processed 20M+ raw records for a global pharma client, then used explainable ML (SHAP) to find the factors driving medicine adoption - Forecast daily fan engagement for a pro sports league from 2M+ records (Kaggle Silver, top 5%) HOW I WORK: - I start with a 4–8 hour batch on your data to measure a baseline - You get the code and a short report with the numbers before and after - Written updates throughout the project - No promised target accuracy before I've seen your data To start, send a sample of your data (a few images, a short video, some documents, or a data extract) and tell me what you need. KEYWORDS: AI engineer, machine learning, computer vision, YOLO, object detection, image segmentation, medical imaging, edge AI, NVIDIA Jetson, LLM, RAG, generative AI, vision-language model, LLM fine-tuning, data science, forecasting, predictive modeling, data visualization, PyTorch, OpenCV, Python, n8n, Power Automate, AI video editing, NLP