Browse categories
Explore
Fiverr Pro
English
$
USD
I will build computer vision for image and video object detection and tracking
Machine Learning Engineer, AI Agents, MCP and Computer Vision
I build AI agents, MCP servers, computer vision models and document automation pipelines. I have 7+ years of experience in data analysis, automation and predictive systems. My work includes COCO-Val s...
About this Gig
I build computer vision models for images and video, including object detection, tracking, classification, segmentation and pose estimation.
Depending on the agreed scope, I prepare data, train or adapt a PyTorch model, evaluate it on held-out samples, and provide inference code or a FastAPI endpoint.
Relevant work includes weakly supervised segmentation with 53.31% COCO-Val mIoU, skin lesion classification and calibration, and MyShot golf video analysis using 2D-to-3D pose estimation.
We define inputs, labels, metrics, runtime and documentation before work starts. Accuracy and speed depend on your data and hardware; benchmark results are not guarantees for a new dataset.
Please share sample images or clips, target objects, available annotations and deployment requirements.
My Portfolio
Other Data Science & ML Services I Offer
FAQ
What relevant computer vision experience do you have?
My work includes weakly supervised segmentation at 53.31% COCO-Val mIoU, ISIC skin lesion classification and calibration, and MyShot golf video pose estimation. I also build inference pipelines and APIs. We will evaluate your task on representative held-out data.
What if my dataset is small or low quality?
I first review representative samples, annotations and class balance. Depending on the task, transfer learning, augmentation or weak supervision may help. I will explain data limitations and evaluation needs before confirming scope.
Can you help beyond just the CV model?
The agreed scope can include preprocessing, inference scripts, a FastAPI endpoint and deployment instructions. Please specify your runtime and integration needs before ordering. Hosting, hardware and ongoing operations are separate unless included.
How do you handle projects that don't reach target accuracy?
We agree on the evaluation split, metrics and deliverables before development. If a target is not met, I report the measured results and error analysis, and discuss improvements within the agreed scope. I do not guarantee accuracy before examining your data.
