Hi fellow data professional!

I’ve broken my own data project rule. I’ve used the same data over and over again.

For 3 years.

It sounds boring but that depth exposure may actually be one of the few moats that slows encroaching AI.

A little context: I support subscriptions, newsletters and growth for my employer. Spoiler alert: These areas are all basically the same thing. And they use basically the same three data sets.

While I have opportunities to jump to other projects, this has been my “area of focus” since I initially led a data platform migration to new vendors in 2023.

So I’m not exaggerating when I say I’ve been staring at the same data for the last 3 years.

Don’t get me wrong. This can get monotonous. But being embedded as the supporting engineer for a particular business area actually allows me to grow my influence internally. It’s allowed me to be considered a “go-to” on any data request relating to this business area.

I realize you may not be currently in a data role or on a team so if you take anything from my experience understand this: To distinguish yourself within an industry that will be flooded with intelligent solutions and contract labor you need to cultivate the ability to go deep on a problem.

I mean like you could dedicate years of your professional life to extracting value deep.

If you’re someone who has spent time building out complex projects with multiple data sources I encourage you to build your next project using just one robust source of data.

• Focus narrowly on use cases, (in your presentation)
• Identify hypothetical stakeholders or downstream users
• Articulate how you’d respond to requests using the data

Out of these, the last point is critical. Corporate leaders are increasingly subscribing to the belief that LLMs can be trained on existing code and API docs and spit out a pipeline. But this is a broad implementation that likely

• Won’t scale
• Won’t address business needs

To be marketable as a candidate, you don’t just want to show how you can go from A to B (requirements->pipeline). You need to go from A to C (requirements->pipeline->scale/support).

Doing this will make you appear to be the “missing piece” frustrated engineering managers crave in a world of candidates that can’t connect the dots.

And if you manage to make the jump, see how I’d establish ownership to become the de facto team expert on a given platform.

Thanks for ingesting,

-Zach

​Medium | LinkedIn | Ebooks