Datature

Datature is a Singapore-based company developing an end-to-end MLOps platform for building computer vision models without the user having to write code. Datature's cloud-based platform allows teams and enterprises to manage datasets, annotate, generate synthetic data, train and deploy computer vision models.

The platform has expanded with new features to improve the developer experience and is capable of full end-to-end applications, such as labeling to model performance. The company is working to deploy a neural network API.

Datature was founded in December 2019 by Keechin Goh (CEO), Denzel Lee (CTO), and Mackenzie Chng. The company is based in Singapore. Datature has been awarded the grand prize of $60,000 at the Rolls-Royce Data Challenge, was the winner of SFFxSwitch SLINGSHOT 2019 Consumer Startups, Top 12 Global Startups by StartupSG, and Top Rising Startups on Tech in Asia. 2nd Product-of-the-Day @ ProductHunt.

Datature's Nexus platform builds and manages computer vision projects. It allows users to perform a variety of functions:

• Uploading images
• Drawing annotations
• Building model training workflow using a drag-and-drop interface
• Monitoring training progress
• Deploying using a one-liner API to load models into their scripts

Demo of Datature's platform - Training an Object Detection Model

Portal is an open-source visual inspection tool for vision models aiming to replace cv2 or matplotlib codes. Portal allows users to load and visualize their deep neural networks on images and videos.

Intellibrush is an AI-powered annotator aiming to make data labeling simpler.

Datature's products have applications in a range of industries:

• Pharmaceutical and healthcare—automation and acceleration of patient diagnosis and drug discovery
• Retail and e-commerce—product recommendation systems, visual search, shelf analytics, etc.
• Smart city— traffic safety, footfall analysis, security, etc.
• Utilities and energy—automating routine inspections and assessments
• Agriculture—crop disease detection, harvest optimization, land management, etc.
• Manufacturing and construction—defect detection, robotics line processing, remote construction monitoring, etc.

In January 2020, the company raised $150K in pre-seed funding led by an angel investor. Datature has also received investment from Enterprise Singapore, a government agency awarding funds for enterprise development.