When to leave your engineering job

Leave when the rate at which you are learning, earning, or building options has flattened for two or more quarters and you have a concrete reason to believe it will not recover. That is the whole test. Everything else (a bad sprint, a reorg, one frustrating manager) is noise unless it shows up as a sustained flat line in growth, comp, or runway. This post is about reading those three lines and timing the move well, not about how to run the search itself.

If you have already decided and just want timing on the actual search, jump to the best time to job search as a software engineer. Here we stay on the prior question: how to know the line has gone flat, and what to do about it.

What are the real signals it is time to leave?

There are three honest signals and a long list of fake ones. The honest signals are stalled growth, comp drift, and bad runway. Each is measurable. If you cannot put a number on your reason for leaving, it is probably a mood, and moods pass. Wait two weeks and check again.

Stalled growth

Growth has stalled when the work stops changing you. A useful proxy: could the version of you from 12 months ago do your current job with a week of ramp? If yes, you have stopped compounding. Concrete markers:

• You have not touched an unfamiliar system, language, or problem class in two quarters.
• Your scope has not grown. Same blast radius, same decision authority, same review weight you had a year ago.
• You are the most senior person on every thread and no longer learn from anyone above you. Mentorship is a real form of comp, and it has gone to zero.
• Promotion has been "next cycle" for three cycles. Read how to become a staff engineer if the ladder itself is the blocker, not the company.

Comp drift

Comp drift is when your pay falls behind the market while you stand still. The market for senior engineers resets every year, and a 3 percent merit bump does not keep pace with a 2-year-old offer. Check yourself against current numbers in senior software engineer salary, SF and NYC 2026. A worked example, illustrative, not advice:

• You signed at $185,000 base two years ago. Two 3 percent bumps put you at about $196,000 today.
• The current median for your level and city has moved to roughly $215,000.
• That is a $19,000 gap, plus a refresh-grant gap on equity. The fastest way to close it is almost always a move, not a counteroffer.

Equity drift matters too. If your initial grant is mostly vested and refreshers are small or absent, your forward comp is quietly dropping every month even if base holds. See how stock options and vesting work to model what is actually left on the table.

Bad runway

If you are at a startup, runway is the signal that overrides the other two, because it sets a deadline you do not control. Runway is months of cash left at the current burn. When it drops under 9 to 12 months with no round in sight, the job you have is not the job you will have. Hiring freezes, scope cuts, and layoffs cluster in that window.

You can read runway from the outside before it is announced. A round that has gone quiet for 24-plus months, a sudden focus on "efficiency," backfills that never get approved, and a quarter where no new headcount opens are all tells. We wrote the full list in signals a startup is about to hire; run it in reverse and you get the signals a startup is about to stop.

Which reasons are not signals?

Plenty of strong feelings are not exit signals. They are problems to raise with your manager first, because leaving does not fix them and the next place may be worse.

How do you time an exit with the funding cycle?

The single most useful timing lever you have is the funding cycle of the companies you want to join, not the one you are leaving. Hiring at startups is bursty. It spikes in the weeks after a round closes, when the budget is approved and the mandate is fresh, then normalizes. If you time your search to land in those windows, you interview with companies that are funded, urgent, and hiring at level.

This is the entire sorting logic of roles.cc: roles are ordered by how recently the company raised, so the freshest post-raise demand sits at the top. You can watch the raises land in near real time on the recent raises page. For the why behind the signal, see why funding recency is the best hiring signal and how funding cycles affect engineering hiring.

Timing your own exit, in order:

1. 01Do not quit first. Search while employed. Leverage in negotiation comes from not needing the offer, and a paycheck buys you the patience to wait for the right one.
2. 02Start when the target market is hot, not when you are most fed up. Q1 and early Q3 tend to have the most open post-raise roles. The full seasonal picture is in the best time to job search.
3. 03Mind your own vesting cliff. If you are weeks from a one-year cliff or a large tranche, the math of waiting is often worth more than the math of leaving. Model it with how stock options and vesting work.
4. 04Land in a post-raise window. A company that closed in the last 90 days is budgeted to hire and moves fast. That compresses your search and improves the offer.

How do you leave well?

How you leave is remembered longer than how you arrived. The engineering community in San Francisco and New York is small, and your manager today is a reference, a future colleague, or a future founder. Leave so that the next person who calls them gets a clean read.

1. 01Sign before you resign. Get the written offer, with start date and comp confirmed, before you give notice. Nothing is real until it is in writing.
2. 02Give honest notice. Two weeks is the floor. Three to four is gracious if you hold critical context and can swing it.
3. 03Write the handoff before the last week. Document the systems only you understand: runbooks, oncall gotchas, the half-finished migration. This is the artifact people remember.
4. 04Tell your manager first, and in person. No Slack, no surprise from a third party. Keep the reason short and forward-looking. "I am taking a role with more scope" travels better than a list of grievances.
5. 05Skip the dramatic exit. Do not torch the burndown in your last standup. Resist a counteroffer used only to plug a hole you already decided to leave through.

On counteroffers: a counter rarely fixes the underlying signal. If you were leaving for stalled growth, a higher number does not create new scope. Studies and recruiter lore both put the majority of counteroffer acceptors back in the market within a year. Take the counter only if it solves the actual flat line, not just the salary line.

If you cannot put a number on why you are leaving, you are probably leaving a mood, not a job.

A simple decision rule

Run this once a quarter, takes 10 minutes. Score each line as flat or growing.

• Growth: Am I learning things I could not do a year ago? Flat or growing?
• Comp: Am I within 10 percent of market for my level and city? Use the salary post for the number.
• Runway (startup only): More than 12 months of cash, or a round closing? Yes or no?

One flat line is a conversation with your manager. Two flat lines, or a runway under a year, is a search. When you get there, the move is not a leap of faith. Aim it at companies that just raised, where the hiring is real and the offer moves. Start with roles.cc, sorted so the freshest demand is already on top.