A Production Problem (Solved)
When Henry Ford attempted to produce the Model T at a rapid pace and with high quality, he ran into a problem. It was difficult to organize teams of specialized workers to assemble automobiles, and with so many workers needed to scale the process, it was highly inefficient. To make matters worse, late delivery of parts caused pile-ups of workers vying for space to work and delays in production.
Harper’s Weekly speculated in 1910 that the person who solved this problem “will not only grow rich but will be considered a public benefactor.” Indeed, Ford’s vision and leadership brought about innovations that led to the development of the moving assembly line in 1913 and laid the foundation for his successful automobile brand. The basic concept of that model - bringing the work to the workers - transformed the way products were mass produced, creating a production model that is still in use by today’s manufacturers.
The New AI Factory Model Emerges
Fast forward to 2018, and teams developing AI solutions face a similar challenge:
How do you transform data (raw material) through multiple process and review steps to prepare it for machine learning?
A new factory model has emerged: the data production line. If you want to develop a high-performing machine learning model, you need smart people, tools, and process management that work together to process massive pipelines of data with high quality. A data production line that includes quality assurance also reduces the need for rework, or having to revisit the data more times than necessary to resolve problems.
Join Our Webinar on Nov. 14
CloudFactory has worked on 150+ AI workflows, and we want to share some of what we’ve learned. We’ve teamed up with our friends at Labelbox, a leading annotation platform, to show you how to turn your data processing into the new AI factory model — the data production line.
Join our webinar on Wednesday, November 14 (11 a.m. PST/2 p.m. EST) to hear from experts in technology and people operations who are transforming the way data is processed and structured for machine learning algorithms.
Here’s a preview of what you’ll learn on the webinar:
- How rework on your data production line slows progress and stifles innovation
- Why crowdsourcing will cost you more time and money than you think
- Proven methods and tools to simplify your production line and accelerate the building of high-quality data pipelines