I was several iterations into coding my third ATS integration — Workday — when I found an Apify actor that solved the complexity for me.

That may not sound like much. But each iteration involved reading API documentation, handling authentication, parsing response formats, and debugging why the data wasn’t shaped the way I expected. Multiply that by a dozen applicant tracking systems, and I was looking at weeks of work before I could even start on the part I actually wanted to build.

That Apify actor hit the jackpot: over forty ATS platforms including Greenhouse, Lever, Workday, Ashby, and dozens more. Already connected. Already maintained by someone else. With a standard input and output format.

I felt two things simultaneously: relief, and a little embarrassment that I’d started building it myself. This is a lesson I’ve learned before, and apparently needed to learn again: the goal isn’t to build everything. The goal is to build the part that doesn’t exist yet.

The ATS connectors existed. The workflow logic for my specific use case didn’t. The AI evaluation layer didn’t. The daily digest tailored to my job search criteria didn’t. Once I plugged in the Apify node, I went from "building plumbing" to "building the system." The energy shifted, and I started solving my actual problem instead of fighting infrastructure.

It’s tempting, especially with a systems background, to want to build from the ground up — to understand every layer, to own the whole stack. But sometimes the most valuable skill is recognizing what not to build.

Apify costs money. It’s worth every penny. Because it let me focus on the part that actually mattered.

Next up: the discipline that kept me from becoming dependent on code I didn’t understand.