Bonsai and NVIDIA: Lowering the Barriers to AI

by Mark Hammond

Artificial intelligence is booming. There are an estimated 21 million developers in the world who want to use AI to solve a specific problem, but without first needing to configure the algorithms running under the hood. Bonsai is working with NVIDIA to overcome this challenge.

AI is increasingly all around us. It powers Netflix’s movie suggestions. It allows the personal voice assistants in our phones to monitor our schedules. And it is making medical diagnoses with unprecedented speed and accuracy. But, AI can only continue to solve problems of increasing scope and complexity — like putting a network of self-driving cars on real city streets — with stronger computing power and developer tools behind it.

To help every developer and any business use AI technology to address real business problems, Bonsai is building a software platform that takes advantage of NVIDIA’s accelerated compute platform, which lowers the barrier to AI.

Smarter Applications Faster

In traditional programming, the progression typically goes from machine code, to assembly language, to C/C++ and, eventually, to higher level languages and visual programming tools. With each step, the code abstracts away some of the complexity. In turn, each layer of abstraction lets a larger group of developers build more proficient programs in less time.

AI is at the level of the assembly language currently. Toolkits like TensorFlow are phenomenally helpful for data scientists previously used to working at the equivalent of the machine code level, but there are only about 19,000 data scientists worldwide. Bonsai’s AI Engine works at a higher level of abstraction so millions of developers, and the companies that employ them, can more efficiently build AI into applications and systems.

Imagine when the developers at GE, the U.S. Department of Education or the Red Cross are able to program intelligent applications as quickly and collaboratively as they might program a database. Scaled with AI technology, the unique expertise and data locked within these organizations stand to create monumental change. But solving problems of such large scope at a faster pace requires immense computing power.

Powering the Future with GPUs

NVIDIA is leading the industry in deep learning hardware. The recently launched NVIDIA DGX-1 was designed specifically to reduce the training time for increasingly sophisticated deep neural networks so that researchers could solve larger problems more quickly.

DGX-1 chart The most creative and important applications of AI will come from a mix of Fortune 100 enterprises, innovative startups, NGOs and solo developers with entrepreneurial drive. Together, Bonsai’s platform combined with NVIDIA’s AI computing innovations significantly extend the reach and capabilities of AI to power the next wave of innovation globally.