When Should You Actually Leave Your Tech Job?

The 2025 Stack Overflow Developer Survey found that 54.4% of developers are actively exploring other opportunities, yet only 11.6% had actually made a move in the prior year. That gap is the thing worth thinking about. A lot of engineers are running the “should I leave?” calculation almost continuously, and most of them are getting stuck on it.

This isn’t a checklist post that tells you “5 signs it’s time to go.” Those posts get read, bookmarked, and then ignored because they’re too abstract to apply to your actual situation. This is more about the specific things that really do predict whether leaving is the right call, versus the things that feel urgent but aren’t.

The signals that actually matter (versus the ones that feel urgent)

There’s a difference between “I’m unhappy” and “I should leave.” They overlap, but not completely. The Stack Overflow 2025 survey found that only 24.5% of developers report being happy at work, which means three quarters of the industry is operating somewhere between complacent and miserable. That can’t all be “leave immediately” situations. Most of the time, unhappiness is information, not a verdict.

The signals that correlate most strongly with leaving being the right decision tend to be structural, not emotional:

• Your compensation is 20%+ below market and has been for at least a year. A one-time pay gap can be negotiated. A persistent gap after you’ve tried is a sign the company doesn’t value the role the way you’d need them to. Check against BLS salary data and levels.fyi for your level and location.
• You haven’t learned anything you’d put on a resume in 9 months. Six months is sometimes a plateau. Nine months is a pattern. If a recruiter called you today, would the last year of your work make you more interesting to them?
• Your manager can’t or won’t sponsor you for the next level. Feedback is cheap. Sponsorship is what actually moves your career. If your manager doesn’t know your work well enough to advocate for it, or doesn’t have the influence to do so, that’s a ceiling, not a floor.
• The company is in financial trouble and leadership is being vague about it. You can read hiring freezes, sudden departures of senior engineers, and shifts from “building” to “maintaining existing customers” pretty clearly. Act on that reading.

There are also things that feel urgent but are usually fixable. A bad project assignment, a frustrating sprint, a manager who’s checked out but is probably leaving themselves. Leaving over any single one of these, without trying to change it first, is usually premature.

When to stay even if you’re miserable

Under 18 months of tenure requires a very strong reason to leave. Not impossible, but you’ll spend 15 minutes of every subsequent interview explaining it. If the reason is “my manager micromanaged me” or “the tech stack was old,” you won’t get much sympathy. If the reason is “the company ran out of money” or “I witnessed an ethics violation I documented,” that’s different.

Upcoming equity vests change the math in ways that are often underestimated. If you’re 5 months from a cliff vest or an anniversary grant, the expected value of waiting is almost always positive, unless your company’s stock is heading toward zero. I don’t know how to tell people exactly where that threshold is, because it depends on the tax situation, the strike price, and your personal risk tolerance. But it’s worth modeling it out rather than leaving it unexamined.

And if you’re currently the best engineer on your team, surrounded by people you’re learning from, solving problems that are hard in ways that interest you, those conditions are rarer than they feel when you’re inside them. The average engineering tenure at tech companies is around 2 years 11 months, according to Ravio’s 2025 compensation research. Engineers move. But “everyone’s doing it” isn’t a career strategy.

The health and ethics cases are in a separate category

Two situations where I’d say leave without needing to fully optimize the timing: when your health is clearly degrading from the stress, and when you’re being asked to participate in something ethically wrong.

On health: chronic sleep disruption, anxiety that bleeds into weekends, physical symptoms that your doctor attributes to work stress. These aren’t “I’m burnt out from a tough quarter.” They’re your body running the numbers ahead of your conscious mind. Delayed career mobility is reversible. Serious health damage often isn’t.

On ethics: if you’ve raised a concern internally, through the proper channels, and nothing changed, you’ve done what you can. Staying at that point isn’t loyalty; it’s exposure.

The mechanics of leaving well

Don’t quit before you have an offer. This sounds obvious. It’s ignored more often than you’d expect, usually because the current situation feels so bad that leaving without a plan feels like the only way to escape it. The risk isn’t just financial. The negotiating position you have as an employed candidate versus an unemployed one is meaningfully different.

The two-week standard notice period is genuinely a floor, not a ceiling. For complex systems or critical teams, offering 4 weeks when you have the offer secured costs you very little and leaves your reputation intact. Tech is smaller than it looks. The senior engineer you handed off to might be your hiring manager somewhere in 4 years.

Write up your systems before you leave. Not to be nice to the company. Because the act of writing it forces you to think about what you built, and that narrative becomes interview material. “Here’s the system I owned, here’s what I documented, here’s what it does” is a much stronger story than “I worked on the payments team.”

The question nobody asks but probably should

Most people frame this as “should I leave my current job?” A more useful version is: “what do I need to be true about my next job that isn’t true about this one?” If you can’t answer that concretely, leaving just moves the same underlying problem to a new address.

The BLS projects 129,200 annual openings for software developers through 2034, and the long-run growth rate for the profession is still 15% over the decade. The market for good engineers is not closed. But “the market is fine” doesn’t make a specific move at a specific time the right one.

If you’re in the “considering strongly” bucket from the Stack Overflow data (about 14.8% of developers), the most useful thing you can do right now is find out what your skills are actually worth on the open market, by doing the interview loop before you commit to anything. That information costs you a few hours. It might save you a bad decision in either direction.

You might leave sooner than you planned, or you might realize the grass isn’t greener yet. Either outcome is more useful than running the “should I leave?” loop indefinitely in your head.

For prepping that interview process, LastRoundAI’s mock interview tool lets you run realistic loops for the specific companies and roles you’re targeting. And if you want to scope out what the comp packages actually look like at those companies, this breakdown of FAANG compensation in 2026 is worth reading before you evaluate any offer. Once you’re deep in the process, sharpening your behavioral interview answers is usually the piece that gets underestimated.