(no title)
mydriasis | 1 year ago
It sounds like you need a couple of people, because you have a couple of problems. The programmer who
1. Understands machine learning and data science, and can action-ably put code in place that utilizes all of the data you're collecting 2. Understands the value of testing and will fix broken windows as he goes ( your "things break constantly" problem ) 3. Can write high-quality code 4. Can move at startup speed while doing all of the above 5. Can spin up at your workplace in time to make an impact around your nearest deadlines of... two or three months from now
surely exists!
But they sound very expensive.
Why:
In my experience, data scientists and machine learning folk are very scientist. Especially at the master+ level, they can make a huge impact on their area of expertise for sure! The drawback -- _usually_, and I will say I'm painting broad strokes here based on a lot of the people I've met and worked with -- is that they're not used to building production software and keeping up with demands of testing and high code quality.
On the other hand, enterprise developers are very good at the high code quality and testing part, but generally don't have the scientific depth to jump in on machine learning / data science projects. Again, broad strokes.
If you want someone who can do both, they usually have a _ton_ of years of experience, because! It takes a ton of years to learn how to do these things well.
So that person probably exists, they're probably quite expensive, though.
If you do wind up with more than one person and you cannot hire them at the same time, you will sacrifice in one aspect for gain in the other. If you hire one person, you run the risk of dragging them through the dirt, because you're going to be stretching their talent thin. I've noticed that the type of person with the level of experience to do the jobs you want don't like to move fast and run themselves ragged.
All this to say --
it sounds like you have lots to do. Be careful!
No comments yet.