Computer Vision Model for Waste Detection and Analytics

Job ID: 39967479

Budget: $30 – $250 USD

Build a Computer Vision Model for Waste Detection (YOLO / VertexAI / Roboflow)

Project Description

We’re developing a waste intelligence and analytics dashboard and need a computer vision expert to train an object detection model on our collected images.

We currently have 100–300 labeled waste images (plastics, paper, e-waste, metal, etc.) and are looking for someone who can quickly train and optimize a detection model that can:

Identify and classify different types of waste accurately

Provide confidence scores for each detection

Deliver results that can integrate later into a dashboard or reporting layer

The ideal freelancer should have hands-on experience with YOLOv8, Roboflow, OpenAI Vision APIs, or Google VertexAI (any of these is fine — whichever gives the most accurate results).

Deliverables

Trained and tested object detection model

Model file(s) + inference code or deployment-ready endpoint

Confidence score outputs for each class

Ability to export data into Excel/data warehouse platform using API's

Brief documentation or walkthrough on how to reuse/retrain the model later

Budget & Timeline

Budget: $50 (fixed)

Timeline: 7 days from the date we provide the dataset

We’re not looking for anything over-engineered — just a clean, functional MVP-level model that helps us analyze waste categories visually.

Bonus Points

Prior experience working with waste recognition or sustainability projects

Ability to explain model performance metrics simply (precision, recall, F1)

Familiarity with exporting to platforms like Roboflow or VertexAI for scalable deployment

Our Goal

We believe in clear communication and quick collaboration — this is meant to be a win-win pilot project.

How to Apply

Please share:

A few lines about your past work in computer vision / object detection.

The tech stack or platform you’d prefer (YOLO, Roboflow, VertexAI, etc.).

Estimated accuracy or validation plan you’d follow for the dataset.