Microsoft
Project Bonsai

bonsai_hero@2x-2

Fresh integrated a Machine Teaching platform that simplifies and optimizes the process of building autonomous industrial control systems, giving users the ability to easily cultivate AI solutions

The Challenge

A leading manufacturing company needed to address three ongoing challenges at their plant: skilled labor shortages, the long lag time in training novice machinists, and the need to get the most life out of costly machines. One process improvement that the client considered was designing and implementing software to automate the creation of CNC machine code, but they felt unsatisfied with the options on the market.

Our Solution

We partnered with Microsoft and stepped in to provide AI and Machine Learning expertise and integrate their cutting edge solution: Project Bonsai. Bonsai learns how to operate a CNC machine efficiently and gives programmers recommendations for new part parameters. An automated workflow aligned with the existing approach to new part creation and helped reduce the expertise and time needed to introduce machine code efficiently and safely.

Key Contributors

Sean McKay
Sean McKay
Kevin Ulrich
Kevin Ulrich
Zach Preston
Zach Preston
Li Ding
Li Ding
Avinash Singh
Avinash Singh
Scotty Paton
Scotty Paton
Mitch Tolson
Mitch Tolson
Darshit Dalal
Darshit Dalal
Aaron Hawkins
Aaron Hawkins
Logan Bean
Logan Bean
Matthew MacKay
Matthew MacKay
bonsai_fifty-fifty_pathway-to-ai

A pathway into AI

Project Bonsai helps engineers with limited data science and coding backgrounds to create working AI solutions. Every feature of the tool optimizes the process, with extensive, built-in validation of a platform in a simulated environment before it’s deployed to a live setting.

bonsai_fifty-fifty_built-to-scale

Built to scale and extend across projects

Creating AI solutions is valuable; making them easily extensible makes that value exponential. Through features built into the Bonsai engine, operators can curate an AI knowledge base with a reward center and goal setting, making the parameters available to all engineers on a team. Bonsai is built so that teams can create one instance and replicate AI solutions easily across multiple products, creating unprecedented levels of scalability. 

bonsai_fifty-fifty_full-control

Full control over what you build

Project Bonsai supports AI designs that work collaboratively alongside people or as independently operating systems. The key is that the platform allows teams to customize, design AI tailored to their infrastructure rather than being one rote solution. 

Moab: Powered by Project Bonsai

We created Moab, a ball-balancing robot, to illustrate the power of the platform. Moab is operated using motion control, visualized in simulation on Project Bonsai, coordinated using trained Bonsai brains, and then deployed to the physical bot. It’s an educational tool and a showcase for the value that Bonsai can bring organizations, with hundreds of possible use cases across various industry verticals.

bonsai_fifty-fifty_cross-collaboration

Cross-disciplinary collaboration

Almost every aspect of our end-to-end services is represented with Moab. Our team built Moab to complete the initial objective of balancing a ball via machine learning, representing disciplines across strategy, design, development, and engineering. The solution allows engineers to experiment with AI and ML paradigms, teaching the robot to catch balls thrown toward it or even after they bounce on a table. The tool provides a playground for mastering intelligent control systems.

bonsai_full-width_a-way-forward-2
bonsai_fifty-fifty_cross-collaboration-2

A platform for increased efficiency and profitability

If a company can shave a few minutes off a process, that’s an opportunity to improve profitability. Empowering machinists to do their jobs better makes good financial sense. 

Project Bonsai has elements of education, extensibility, and knowledge sharing, but ultimately it comes down to delivering value across various industries. For the leading manufacturing company we partnered with, we provided a solution that addressed and optimized their razor-thin margins. Most importantly, it’s a solution that they can continue building on, teaching with, and improving with their base of knowledge and expertise.