Skip to content

Filling Out Lever Job Applications — Plain-Language Notes

WorkWingman applies to jobs for you. A lot of companies run their application forms through a system called Lever (the web address looks like jobs.lever.co/somecompany/…/apply). To learn how to fill those forms reliably — without ever touching a real one — we built a little practice range.

The practice range (and why it's safe)

We built a fake Lever form that only lives on this computer. It runs on 127.0.0.1, which is the address a computer uses to talk to itself. Nothing goes out to the internet. There is:

  • no real Lever website involved,
  • no account created anywhere,
  • no Submit that does anything — the Submit button is a decoy that just writes a note in the log ("would submit — suppressed") and starts over.

Then we let the robot practice on this fake form hundreds of times, each time with the fields shuffled and renamed a little, and we wrote down what worked and what didn't.

What a Lever form is like

Think of a Lever application as one long page you scroll down and fill in, top to bottom:

  1. Your name, email, phone, and current company.
  2. A place to attach your resume (in the practice range we never actually attach a file).
  3. A few link boxes — LinkedIn, GitHub, portfolio, and so on.
  4. Some custom questions the specific company added, like "Are you allowed to work in the US?" or "What salary are you looking for?"
  5. Sometimes a consent checkbox (agreeing they can process your info).
  6. A Submit button at the bottom.

The important thing: it's all on one page. You don't click "Next" through a series of steps.

How this differs from Workday (the other big one)

We built the same kind of practice range for Workday, and Lever is noticeably different:

  • Workday is a stack of pages you click "Next" through (sign in, then your info, then work history, then education…). Lever is one single page. Much simpler to fill.
  • Workday makes you log in. Lever doesn't — the public apply page is right there.
  • They label their form boxes differently under the hood. Workday tags each box with a long descriptive code; Lever uses short plain names like name and email, and a distinctive style for link boxes (urls[LinkedIn]). So the robot has to look for different clues on each site.
  • Lever's custom questions use simple dropdowns, which are easy to pick from. Workday uses fancier custom menus that take more steps to open and choose.

How the robot finds each box (and stays reliable)

Companies constantly re-skin their forms, so the robot never relies on a single clue. For every box it has a ranked list of ways to find it — the most reliable first, then backups. If the company renamed something, it slides down to a backup clue and still finds it. We keep a running tally of how often it needs a backup, which tells us the backups are pulling their weight.

Across a practice run, the robot found and filled every form, every time, with zero crashes — even when boxes were renamed, hidden, or slow to appear.

When the robot stops and asks

Some questions can't be answered safely from your saved profile — like "what's your desired salary?" or "how did you hear about us?" For those, the robot does not guess. It pauses and flags the question for a human to answer. In the practice run it flagged 18 such questions across 20 applications. Questions it can answer confidently (work authorization, education level) it fills in on its own.

A real bug the practice range caught

The first time we ran it, something looked wrong: the robot said it had paused for those tricky questions, but the counter said it paused zero times. Digging in, we found the robot was correctly pausing — but a later step (finding the Submit button) was accidentally erasing the pause and marking the application "ready to submit." That's a dangerous little bug: it could turn a "wait, ask a human" into "go ahead." We fixed it so nothing can override a pause, and re-ran — now it correctly reports all 18 pauses. That's exactly why the practice range exists.

The bottom line

  • The robot practices on a fake, local, offline Lever form only.
  • It never submits a real application and never talks to the internet.
  • It learned how Lever forms are laid out, how they differ from Workday, and how to find each box even when the form changes — and it caught a real safety bug along the way.