Projects Blog Tools Music Contact
← All Posts
Startup June 29, 2026

Three Failed Jobs in a Decade: Rebuilding a Dev Career

By: Evgeny Padezhnov

Illustration for: Three Failed Jobs in a Decade: Rebuilding a Dev Career

A decade in software can produce three dead-end jobs in a row. A scrapped project. A worthless equity grant. A funding "pause" that ends a role overnight. The pattern feels personal. It usually is not.

This is a recurring story on Hacker News and in engineering circles. A developer does the work, ships the features, and watches the business collapse anyway. The code was fine. The decisions above it were not. Recovery starts with separating the two.

The pattern, not the person

Read the typical failure sequence carefully. Three jobs, roughly a decade, three different reasons for the exit.

First job: a medium corporation. The exit cause was friction — slow projects, low motivation around, heavy process. Common and survivable.

Second job: a small company. A next-gen e-commerce build with real-time stock checks and multi-currency support. Eighteen months of work. The result: unpaid, then scrapped for a worse site that barely completes a checkout. The engineer shipped. The business chose otherwise.

Third job: a startup. A 3% stake, full ownership of the tech stack, a marketed "$__M valuation, well funded." Eighteen more months. Dozens of features. A payment system built twice because founders changed their minds. Then a "pause" because funding dried up, plus an offer to be rented out as a consultant.

Key point: none of these failures was caused by the engineer's code. Each was caused by business risk the engineer absorbed without compensation. That distinction is the whole article.

What actually went wrong

Three failures, three different root causes. Naming them prevents repeating them.

The scrapped project

Eighteen months of work, unpaid, then discarded. The lesson is about who controls the decision and who carries the risk. A salaried engineer carries time risk. If the company does not pay, the salary should still arrive. If it does not, that is a different problem — a failing employer, not a failing developer.

Common mistake: tying personal pride to a project's survival. Projects get killed for reasons unrelated to quality — budget, politics, a cheaper vendor, a founder's whim. The code shipped. The portfolio entry is real. The death of the project is a business event.

The 3% stake

Three percent sounded large. In plain terms: 3% of an unfunded, illiquid startup is usually worth zero. Equity only converts to money in an exit — acquisition or IPO. Most startups never reach either.

CB Insights analyzed post-mortems and found "ran out of cash or failed to raise new capital" near the top reason startups fail — roughly 38% in their Top Reasons Startups Fail report. A 3% stake in a company that runs out of cash is 3% of nothing. The "pause" in the third job is exactly this failure mode.

Building the payment system twice

Founders who do not know what they want will pay for the same work twice — in the engineer's hours, not their own. Rebuilding a payment system because requirements flipped is a management failure. It looks like engineering churn from the inside. It is indecision at the top.

Equity in plain terms

Most developer pain around startups comes from misunderstanding equity. Three numbers matter, and a stake percentage is not enough.

Valuation and dilution. A 3% stake today shrinks with every funding round. After two more rounds it might be 1.5%. Future dilution is rarely explained at signing.

Strike price and vesting. Options are not shares. They are the right to buy shares at a fixed price, earned over time — typically four years with a one-year cliff. Leaving early forfeits unvested options.

Liquidity. Private shares cannot be sold freely. Money appears only at an exit. No exit, no money.

Try it: before accepting any equity offer, ask three questions in writing. What is the current fully-diluted share count? What is the strike price? What is the most recent preferred-share price? If the answers are vague, treat the equity as worth zero and judge the cash salary alone. The Holloway Guide to Equity Compensation covers the mechanics in depth.

Key point: equity is a lottery ticket, not a salary. Accept a startup role only if the cash compensation alone is acceptable. The stake is upside, never the basis.

Due diligence before joining

The third job was "marketed" as well funded. Marketing is not verification. A developer can check funding the same way an investor would, in under an hour.

Check the funding record. Crunchbase and PitchBook list rounds, amounts, and dates. A company claiming strong funding with no record since two years ago is burning down. No recent round plus a growing headcount equals a coming "pause."

Estimate the runway. Ask directly: how many months of cash remain at the current burn rate? A serious founder knows the number. A vague answer is the answer. Anything under twelve months is a red flag for a new hire.

Read the cap table signals. Ask how much the founders raised and at what valuation. Ask how many employees the company added in the last year. Rapid hiring on flat funding shortens runway fast.

Tested in production: candidates who ask runway questions in interviews get sharper, faster offers from healthy companies — and visible discomfort from unhealthy ones. The discomfort is the signal.

Common mistake: trusting a verbal "$__M valuation." Valuation is set by the last investor, not the founder's pitch. Unconfirmed numbers are noise.

The financial buffer changes everything

A "pause" is survivable with savings. It is a crisis without them. The buffer is not just safety. It is negotiating power.

Standard advice across personal-finance and engineering communities: keep three to six months of expenses in cash. For developers in volatile startup roles, six months is the floor, not the ceiling.

In practice, the buffer does three things. It removes the pressure to accept the next bad job immediately. It allows refusing a "rent me out as a consultant" demand. It funds the gap between leaving and landing a role that is actually good.

Key point: the engineer who can say no for six months gets better outcomes than the one who must say yes tomorrow. Cash buys that no.

Getting up after the third failure

Resilience is concrete, not motivational. Specific moves restart a career after a streak of bad jobs.

Separate signal from story

Three failures feel like a verdict on ability. They are not. The code shipped in all three jobs. The e-commerce platform worked. The payment system worked — twice. The features worked. In plain terms: the engineering was sound; the employers were not. If it works — it is correct. The output was correct each time.

Extract the real portfolio

Every failed job produced shippable evidence. Real-time stock checks. Multi-currency support. Two production payment integrations. Dozens of features. That is a strong résumé regardless of whether the companies survived. Document the systems built, the scale handled, the problems solved. Dead companies still produced live skills.

Change the selection criteria

Three failures share a theme: weak or dishonest businesses. The fix is not better engineering. It is better employer selection. Screen for funding stage, runway, founder track record, and revenue. Prefer profitable or well-funded companies over "fast growing" pitches. The next job's quality depends on screening, not skill.

Decide on the next bet deliberately

After a startup "pause," the instinct is to grab anything. Resist it if the buffer allows. Three paths exist, each valid:

Common mistake: jumping straight into the same kind of company that just failed. Repeating the input repeats the output.

What to try right now

Pick one job from the last three. Write down every system shipped, the scale it handled, and the concrete problem it solved. Strip out the company's outcome entirely. That document is the real résumé — proof the engineering held even when the business did not. It also reveals which skills to lead with in the next search.

The failures were real. They were also mostly not the developer's. Recovery starts by naming which risks to stop absorbing for free — unpaid project risk, worthless-equity risk, no-runway risk — and refusing to carry them again.

Information is accurate as of the publication date. Terms, prices, and regulations may change — verify with relevant professionals.

✦ Squeeze AI
  1. Все три провала произошли не из-за кода инженера, а из-за бизнес-рисков, которые он поглощал без компенсации — это разграничение и есть суть статьи.
  2. 3% в неликвидном стартапе без экзита (M&A или IPO) стоят ноль: по данным CB Insights, ~38% стартапов гибнут именно потому, что заканчиваются деньги.
  3. Фаундеры, которые не знают чего хотят, платят за переделку часами инженера, а не своими — двойная сборка платёжной системы это не технический провал, а сигнал о проблеме принятия решений выше по цепочке.

Powered by B1KEY