Transforming Open Source AI into Live Service REVIEW

The Open Source movement, proposed by Richard Stallman in the late 1980s, has become the most important community for interacting and improving programs freely. It is also playing a crucial role in AI development; more than 80% of the development process is based on open source. However, when trying to make use of open source, developers experience various hurdles such as difficulty in understanding code, build failure, lack of suitable resource environment, etc. To solve these problems, we are working on converting an open source project into a live service and making it always available as an API. Once projects are always executable, anyone can utilize them right away, and anyone can combine open services to propose new services. Currently, more than 400 open source projects have been deployed as an open service, and our goal is to enable open source developers to use the distributed service as an API through an automated pipeline. As a result of running an automated pipeline targeting Github so far, only 7.5% of them are containerized and serviceable. In particular, AI solutions are considered to be difficult to service due to complex cloud and dependent variable settings. We want to introduce you to some of the best practices of deploying open source projects as a service with live demos using containers and container orchestration. Finally, I would like to propose the ideal ecosystem in which multiple open source projects cooperate on the open cloud, a.k.a Open Resource
  • Minhyun Kim
  • Minhyun Kim / CommonComputer CEO
  • Minhyun has been developing search engines, machine learning, and user models for about 8 years at Google, and has worked as a technical advisor for the Google Machine Learning Crash Course and Google Impact Challenge. For many years, he has participated in various developer events such as OpenHack and TensorflowKR ML Camp as a mentor. To solve difficulties experienced while developing AI with open source developers, he is currently developing AI Network at Common Computer. AI Network is a federated serverless cloud service that connects open source AI solutions.

Feel free to leave comments/curiosity/opinions about the session.

* Comments (including swearwords and slander) that do not match the purpose of the bulletin are deleted by the administrator without notice.

* Comments can only be uploaded while the sessions are ON–AIR and updated every minutes.

0/100 characters WRITE