🤓[DS Career] Interviewing? How to pick the right job
Trying to get a data science job is as risky as launching to prod with no monitoring...
Make sure the incentives of the business support the job description.
I’ve made this mistake a few times now in my career. I chose a prestigious-sounding job that ended up being bad for my career.
To be clear, “bad” is a relative term. Of course I learned good things from these two companies. I developed my skillset and became more well rounded as a result. But I wasn’t spending those three years doing the work I really, truly enjoyed.
On my most recent job, my non-technical hiring manager sold me on a great pitch: “we want you to come harvest all of our data, build models, and make our product better.” Nothing sounded more appealing: ML-driven products. Yum.
But once I started working I realized something: the business doesn’t need optimal solutions. They needed quick-and-dirty, good-enough SQL queries. And so, I became frustrated because I wanted to optimize but they wanted something sub-optimal.
How can you identify this during the interview process?
Look at the incentives of the business. How does this business make money? If it’s a big bank, they make money on underwriting. They process billions of transactions and live and die based on expected values. They need optimal solutions.
If you’re looking at an ecommerce business, it’s likely they’ll want you to help on their pricing. But perhaps they don’t actually control the prices because they sell third-party inventory. The way your ecommerce business makes money is by getting better volume deals, not necessarily by optimizing prices. They may even be in price-lock contracts. So you end up doing business analytics.
I recently was hired by a company who’s main marketing pitch is how accurate their models are. Okay, the incentives are aligned. It’s been great since day one.
So when looking for a job, make sure the incentives are aligned. Make sure they depend on optimal solutions to move the business forward.