The current and upcoming needs of businesses will include machine learning.
All needed are a dataset and a desire for an accurate prediction, some distilled understanding, or repeatable content creation. Great examples include:
- What will be the most likely value of a business metric in the future?
- What’s the most likely price of a dozen eggs next week?
- How likely is it that someone does some action?
- What important information is in pictures, videos, or text?
- Can you translate this language into another?
- Machine Learning can do that.
The opportunities are there now for businesses, and even more, they will exist in the future.
While tractable ML opportunities DO exist now, there’s no easy way for most individuals in a company to apply machine learning to their data and get a repeatable, reliable, business-changing result.
It’s easy to find a how-to guide for the math or put pre-trained models into a micro-app. Not so easy for actual production.
There aren’t many companies that have used cutting-edge machine learning day in and day out to power their business at scale; there’s an opportunity to lead the market with an ML-friendly platform. Large parts of most companies’ infrastructure aren’t ready to even attempt. There’s a wide gap between where many current companies are (data in excel, unstable infrastructure, database with no easy access to exploration, no monitoring) and the proven run-away success seen by ML pioneer companies. We should guide our customers towards proven best practices like easily repeatable experimentation notebooks, automatically re-trainable models, pulling data for ML from reliable infrastructure, all monitored, operated, and deployed automatically.