Machine Learning for Business

Machine Learning and Artificial Intelligence have been buzzwords for a few years now. Up until only recently though, a company had to spend big and have in-house expertise to have any kind of feasible and useful AI adoption.

What changed? The ability to implement AI models in production environments went from requiring a team of PhD data scientists and software engineers to just a handful of skilled Machine Learning Engineers - that’s our term for it anyway. What is an ML Engineer? Someone who can take data, create a useful predictive model*, and deploy it in cloud production environments.

How does this happen? What happened for the need for PhDs to create custom models? The answer is, frankly, the tools out there are just at the point now where one person can be as powerful as 5 used to be. Open source models that can be fine tuned on a customer’s dataset take a fraction of the time to train and optimize. Production environments are created in the cloud now. The big three - AWS, Azure, and GCP - all have robust deployment systems for hosting ML systems.

What is the sum of these developments? The beginning of wide-spread adoption of AI systems by businesses that don’t have the same resources as the monolithic tech giants. The process by which these business adopt the technology is with an increasing number of machine learning consulting firms such as ours, Standard Data.

This is the natural progression of efficiency - there’s a point at which the benefit of implementing AI is not greater than the cost of hiring a full time team for it, but is greater than the cost of hiring consultants for it; or maybe it’s feasible for a company to hire a team, but the cost in terms of time, distraction, and lack of expertise makes it not worth it.

Incumbent consultancy firms have been around and have expanded their abilities to meet this need, but the time for smaller, more cost effective and focused teams is here.