A Basic Framework For Adding Domain Experts To Data Science Projects
Domain experts and frontline workers are often at the fringes or excluded from Data Science projects. Integrating them into projects is complex. Here's how I've done it.
Do you need to be a Machine Learning expert to use the technology? No. Some very successful use cases come from businesses who train frontline employees on basic analytics and give them AutoML tools. None of them are Data Scientists but they can integrate models into their workflows to improve work products.
The interesting and valuable applications happen a lot faster in the frontline than from the Data Science team. What I mean by that is the idea/project quality is much higher when it’s coming from the people actually doing the work or facing the customers.
I’ve only been running initiatives like this for the last 2 years because AutoML is just now becoming useful. I used to rail against AutoML as dangerous in the hands of the untrained, but I didn’t really understand it or how it should be implemented. Not every project needs rigor if there’s a domain expert in the loop. What I’ve found is that the outcomes are a lot better when the domain expert is also doing the model development.