Neurocle

Neurocle is a South Korean deep learning vision software company founded in 2019 by chief executive Hongsuk Lee and based in Gangdong-gu, Seoul. It targets automated quality inspection in manufacturing, where conventional rule-based machine vision struggles with variable defects and where factories lack in-house deep learning engineers to build and retrain models.

The product line covers the full inspection workflow: Neuro-T is an auto deep learning model trainer with AI auto-labelling and eleven model types, Neuro-T Engine embeds that training capability for on-site MLOps, Neuro-R is the runtime library for deployment on production lines, and Neuro-EDU serves education and workforce training. A generative model creates synthetic defect images where defect data is scarce. The company reports passing 1,000 projects and 700 clients as of 2024 across 40+ countries, with applications in semiconductor, battery, automotive, electronics, metal, food and beverage, pharmaceutical and medical device production.

1. Neuro-T: automated deep learning trainer that optimises architecture and parameters, with AI auto-labelling and eleven model types.

2. Neuro-R: runtime library that deploys trained inspection models on production equipment.

3. Neuro-T Engine: embeds training into the line so plants can run on-site MLOps and retrain models themselves.

4. Synthetic defect generation: a generative model creates defect images when real defect samples are scarce.

5. Track record: 1,000+ projects and 700+ clients as of 2024, across 40+ countries.

6. External recognition: Gartner Cool Vendor in AI (2021) and Vision Systems Design Innovators Awards for five consecutive years (2021-2025).