I Reviewed 31 iOS Portfolios in Tokyo Over 9 Weeks — the 7-Step Filter That Found 2 Hires Without a Recruiter

The role sat open for eleven weeks before we changed anything. Not because Tokyo has no iOS engineers — it has plenty — but because we had written a job description that described about nine people, most of whom were not looking. What follows is what we changed, in the order that mattered.

Step 1 — Decide the Japanese language requirement honestly

This is the single highest-leverage decision in Tokyo mobile hiring and most teams make it by default rather than by analysis.

Business-level Japanese is genuinely required when the engineer gathers requirements directly from Japanese business stakeholders, handles customer escalations, or joins a team where product and design discussion happens exclusively in Japanese.

It is not required when the engineer works from written specifications, the team already has a bilingual product owner, or code review and technical documentation are in English.

Ours was the second case, and we had written the first. Changing that one line took our qualified pool from 9 candidates to 31. No other change in this entire process came close to that effect, and the cost was hiring a bilingual product owner we needed anyway.

If you cannot name the specific weekly meeting that requires business Japanese, you do not need business Japanese.

Step 2 — Write the job description around the app, not the stack

Every iOS job description in Tokyo lists Swift, SwiftUI, UIKit, Combine, Core Data. Candidates cannot distinguish between them, so they filter on compensation and company name instead — which is the worst possible filter for you.

Describe the app. What does it do, how many users, what is the release cadence, what is the current crash-free rate, what is genuinely broken. We wrote that our app had 340,000 monthly active users, shipped every three weeks, had a 99.1% crash-free session rate we wanted above 99.7%, and carried a UIKit codebase we were migrating to SwiftUI screen by screen with no deadline.

That last clause — admitting a half-finished migration — brought us three candidates who said it was why they applied. Engineers who have lived through migrations recognise honesty about them, and engineers who have not are not who you want for that work.

Step 3 — Source outside the domestic job boards

Domestic boards reach Japanese-speaking engineers already engaged in a domestic job search. That is one segment, and for an English-capable role it is not the largest one.

Four channels produced our 31 candidates. Domestic boards gave us 11. English-language channels aimed at international engineers in Japan gave us 9, and these were the strongest on average. App Store credits — finding apps we admired and tracing who built them — gave us 6, and this is the most underused sourcing method in mobile hiring. Open source contributors to Swift libraries we depend on gave us 5.

Two structural notes on the Tokyo market. First, the international engineering community in Japan is substantial, already visa-resident, and dramatically easier to hire than someone relocating. Second, Japan's underlying shortage is demographic — the working-age population is shrinking by roughly half a million people a year — so this pool does not replenish. Treat every year you delay as strictly worse.

Step 4 — Review a shipped app instead of setting a take-home

We had a four-hour take-home. Two senior candidates declined the process specifically because of it, and we were about to lose a third.

We replaced it with a 45-minute conversation: “Show me an app you shipped and walk me through a decision you would now reverse.”

The signal is dramatically better. A take-home measures how someone builds a small greenfield project under artificial constraints — which is not the job. A shipped app reveals how they handled an App Store rejection, a migration they could not finish, a performance problem on a four-year-old device, a dependency that went unmaintained.

What to listen for: do they talk about users or about code? Strong iOS engineers frame technical decisions in terms of app size, battery, launch time and older-device behaviour. Weaker ones frame them in terms of architectural purity. Both write working apps; only one of them ships well on the long tail of devices that Japanese users actually keep.

If a candidate has only worked on closed internal apps, do the same walkthrough verbally with screen sharing. The point is real constraints, not public code.

Step 5 — Run a 60-minute architecture conversation on your real codebase

Present a genuine constraint from your app. Ours was: 140 UIKit screens, a partial SwiftUI migration, a team of four, and a product roadmap that does not pause.

Ask how they would proceed. You are not grading the plan — you are grading whether they ask what the business cost of the current state is before proposing anything. An engineer who proposes a full rewrite without asking about release cadence or team size is telling you how the next two years will go.

The follow-up that separated our finalists: “Six months in, the migration is 40% done and product wants to pause it for a quarter to ship a major feature. What do you do?” Strong answers engage with the cost of a half-migrated codebase, which is the actual problem, rather than defending the migration on principle.

Step 6 — Benchmark against 2026 Tokyo bands and confirm the visa route

English-only roles at foreign-affiliated companies sit at the upper end of each band because the pool is smaller. Engineers combining iOS depth with backend capability or on-device machine learning command a further premium — consistent with the broader Tokyo bifurcation, where specialised profiles hold very low effective unemployment while generalist roles do not.

On visas: for candidates already in Japan with a valid status, this step is administrative. For relocations, confirm the route before the offer conversation. Certificate of Eligibility processing commonly runs one to three months. The Highly Skilled Professional points system and the J-Skip route for high-salary professionals both offer materially faster handling, and Japan has expanded these categories specifically to attract international technical talent.

Step 7 — Set a 90-day mandate tied to release health

Write it before they start. Ours had three items, all measurable, none requiring org-wide agreement.

Crash-free session rate from 99.1% to 99.6%. CI build time from 22 minutes to under 10. A written, agreed policy on which screens migrate to SwiftUI and which stay in UIKit indefinitely — the last one being the item that stopped the migration from being a permanent low-grade argument.

At 90 days our first hire delivered all three, and the third one turned out to matter most. The migration had been consuming attention for eighteen months precisely because nobody had been empowered to declare parts of it finished.

The 3 mistakes we made, so you can skip them

1. We required business-level Japanese without checking whether we needed it. Eleven weeks of an open role, for a requirement that did not survive ten minutes of scrutiny. Audit this first, always.

2. We kept a take-home too long. It cost us two senior candidates outright and gave worse signal than a conversation about a shipped app. Senior mobile engineers in Tokyo have options and decline processes with unpaid multi-hour exercises.

3. We waited for offer acceptance to look at the visa route. For one relocating candidate this added five weeks we could have run in parallel. Confirm the route during the shortlist stage, not after.

How this compares across the region

Singapore's mobile pool is deeper and moves faster on compensation, particularly with fintech demand — our colleagues at HireDeveloper.sg track that market. Dubai has a thinner mobile pool but faster visa mechanics, covered by HireDeveloper.ae.

See also our role pages for iOS engineers, Swift engineers and React Native engineers in Tokyo, plus our analysis of what the August 19 Japan-Taiwan Summit means for bilingual engineer demand.

Frequently Asked Questions