React Native hiring changed under recruiters' feet between 2024 and 2026, and most job specs haven't noticed. The framework's New Architecture became the default in version 0.76 on 23 October 2024. By 0.82, released in October 2025, the old architecture couldn't even be switched back on. Meanwhile, AI assistants took over the screens, the API calls and the test scaffolding. In 2027, the code is the cheapest part of a React Native hire, which means your screening has to test the skills an AI can't supply for the developer.
The adoption numbers leave little room for doubt. In the Stack Overflow Developer Survey 2025, 84% of respondents said they use or plan to use AI tools, up from 76% a year earlier, and 51% of professional developers use them every day. Building a scrolling list with pull-to-refresh used to be a fair test of competence. Now it's a prompt.
What I use instead, after years of hiring and leading mobile teams, is the beyond-the-code scorecard. Six non-coding skills, each with a question to ask, the sound of a good answer and the red flag that should end the process. It isn't a replacement for a technical test. It's the filter you run before you pay a senior engineer to sit in on one.
UK React Native roles are scarcer than in 2024 and better paid
The wider UK tech market isn't short of openings. APSCo and CV-Library's technology sector overview counted 126,861 unique tech vacancies in Q2 2026, up 4.3% year on year, with permanent AI Engineer postings up 74.4%. React Native is a smaller and pricier niche inside that. ITJobsWatch lists 176 UK permanent adverts citing React Native in the six months to 9 November 2026, against 147 in the same period of 2025 and 243 in 2024. Over those two years, the median advertised salary climbed from £72,500 to £90,000.
Those are small samples, so treat them as a direction rather than a rate card. The adverts can't prove motive, but here's how I read them: employers now hire fewer junior React Native developers and pay up for people who can own an app from first build to store listing. That matches the briefs I see, where nobody asks for someone to build screens any more. They ask for someone to ship, which brings us to the line every hiring manager eventually says out loud.
A good JavaScript developer can pick up React Native in a week.
It's half right, and the half that's right matters. A strong React developer will be productive on components and state within days, because the JavaScript layer is familiar ground. Stack Overflow's 2024 survey found React Native in use by 9% of professional developers, so the pool of people who already know it is real but not huge, and widening the search to web developers is often sensible.
What takes months is the other half. Why a build passes on a laptop and fails in CI. What Android does to a background task. Why Apple rejected the app over a permission string. The week covers the code, and the code was never the hard part.
If your in-house team can't run that kind of search at speed, outsourcing is a legitimate route. Some companies hire React Native developers through a dedicated-team provider such as LITSLINK and keep product ownership in-house, which works well as long as the vendor's engineers face the same scorecard as your own candidates.
Here's the scorecard on one page. Each row maps to one of the sections below.
|
Skill |
Why it matters in 2027 |
Interview signal |
Red flag
|
|---|---|---|---|
|
Product thinking |
AI builds features fast, so picking the right one matters more |
Talks about users and trade-offs before libraries |
Describes what they built, never why |
|
Native platform awareness |
The New Architecture makes native modules routine work |
Knows when to write Swift or Kotlin and roughly how |
"There's always a package" |
|
Performance profiling |
AI-generated code is rarely tuned for low-end phones |
Names a profiler and a measured before and after |
Fixes speed by memoising everything |
|
Release and store management |
Apple and Google change submission rules every year |
Has paused a rollout and handled a rejection |
Has never opened App Store Connect or Play Console |
|
Reviewing AI-generated code |
Most code now arrives as a suggestion |
Explains what they reject and why |
Can't recall turning a suggestion down |
|
Communication |
Teams span time zones and design tools |
Writes clear async updates, works from Figma specs |
Needs a call for every decision |
Product thinking is the first filter, because AI made building cheap
When an assistant can produce a working settings screen in a minute, the expensive mistake is building the wrong settings screen. Good mobile developers push back on scope. They ask what the analytics say, and they know every extra onboarding step loses users before the app has done anything useful.
-
Ask: "Tell me about a feature you argued against, or cut down, before building it."
-
A strong answer sounds like: a specific user problem, a cheaper alternative they proposed and what happened to the numbers afterwards. Extra credit if they bring up offline states or slow networks without being prompted.
-
Red flag: every story is about a library they introduced. Technology choices are fine. A candidate who can't describe a single user isn't.
Native platform awareness separates mobile developers from web developers on a phone
React Native runs on top of iOS and Android, not instead of them. Under the New Architecture, writing a native module in Swift or Kotlin is ordinary work, and somebody on the team has to know when JavaScript has run out of road. Store rules belong in this skill too. Since 28 April 2026, Apple has required uploads to be built with Xcode 26 and an iOS 26 SDK. Google's target API requirements say new apps and updates must target Android 16 (API level 36) from 31 August 2026.
-
Ask: "When did you last write native code in a React Native project, and why couldn't JavaScript do it?"
-
A strong answer sounds like: a concrete case, such as Bluetooth pairing, background location or a camera feature where the community package fell short. If over-the-air updates come up, they should know Apple's guideline 2.5.2 forbids downloading code that changes an app's features or functionality.
-
Red flag: "There's always a package." Sometimes there is. Sometimes it hasn't been maintained in three years and breaks the build the week before launch.
Profiling is a habit, and most candidates guess instead
Performance complaints rarely arrive with a stack trace. A user says the feed stutters on an older Samsung, and the developer has to prove where the time goes. React Native 0.76 shipped React Native DevTools as its default debugger, and Xcode Instruments, the Android Studio profiler and Perfetto cost nothing. Guessing is a choice. AI-written code makes the problem worse, not better, because it tends to re-render generously and fetch eagerly.
-
Ask: "Walk me through the last performance problem you measured. Which tool, what number before, what number after?"
-
A strong answer sounds like: a named tool, a named metric such as dropped frames, startup time or memory, and a fix like moving heavy work off the JavaScript thread or swapping a list component for Shopify's FlashList.
-
Red flag: "I wrapped everything in useMemo." That's not profiling. It's superstition.
Release management is a skill, and someone on your team has to own it
Shipping a mobile app isn't a deploy button. It's TestFlight builds for testers, closed tracks in Play Console, release notes, review queues and a plan for the day crash rates jump. Apple's phased release spreads an update over seven days, starting with 1% of users and reaching 100% on day seven, and it can be paused for up to 30 days in total, while Google Play's staged rollouts let you set the percentages yourself. Microsoft also retired Visual Studio App Center on 31 March 2025, so teams that used its CodePush service for over-the-air fixes had to move elsewhere. A good candidate will know what their last team chose and why.
-
Ask: "Tell me about a release that went wrong. What did you do in the first hour?"
-
A strong answer sounds like: pausing the phased release or halting the staged rollout, checking crash reports in Sentry or Firebase Crashlytics, shipping a fix and explaining the delay to the people waiting on it. They know which steps need a human in App Store Connect.
-
Red flag: "Our DevOps person handled that." Fine at a large company. At a mid-sized one, it usually means nobody owned it.
Reviewing AI-generated code is now part of the job description
Your candidate will use AI. What you need to know is whether they check its work. The same Stack Overflow survey found 46% of developers distrust the accuracy of AI tools, against 33% who trust them. 66% named "AI solutions that are almost right, but not quite" as their top frustration. Almost right is the dangerous kind of wrong on mobile, where a missing permission prompt or a listener that never unsubscribes only shows up on a real device a week after release.
-
Ask: "Describe a suggestion from Copilot, Cursor or ChatGPT that you rejected. What was wrong with it?"
-
A strong answer sounds like: specifics. A deprecated API, an invented prop, an effect that subscribed without cleaning up, a test that asserted nothing at all. They'll also tell you how they changed their prompts afterwards.
-
Red flag: they can't remember rejecting anything. Either they don't review, or they don't use the tools. In 2027 both are problems.
Communication across time zones and with designers decides whether remote hires work
Most React Native teams I've led were split across at least two time zones, with designers in a third. A developer who writes a clear pull request description, flags a Figma frame that won't fit small Android screens and posts a two-line update before logging off saves everyone hours each week. One who waits for the next call quietly costs the team a day.
-
Ask: "A designer sends you a screen that won't fit on a small Android phone. You're six hours ahead of them. What do you do?"
-
A strong answer sounds like: build the version that works, leave a written note with screenshots, propose an alternative and keep going. Strong candidates also point out where iOS and Android conventions differ, like back navigation or bottom sheets.
-
Red flag: "I'd wait for them to come online." With a two-hour overlap, waiting is a decision, and it has a cost.
A 30-minute screening script
Recruiters can run nearly all of this without an engineer on the call. Score each segment from 1 to 4 and move forward anyone who scores 3 or higher on at least four of the six skills.
-
0 to 3 minutes: set the frame. Tell the candidate the call is about how they work, not syntax. People answer more honestly when they know a whiteboard isn't coming.
-
3 to 8 minutes: product thinking. Ask the feature-they-argued-against question. Listen for a user, a trade-off and an outcome, and note whether they mention a metric without prompting.
-
8 to 13 minutes: native platforms and store rules. Ask about their last native module and the last store rejection or policy change they dealt with. Candidates who bring up API level 36 or Xcode 26 unprompted are current.
-
13 to 18 minutes: profiling. Ask for the tool, the number before and the number after. No numbers means no score above 2.
-
18 to 23 minutes: release ownership. Ask about the release that went wrong. You're listening for four things: a paused rollout, a crash dashboard, a fix and a message to stakeholders.
-
23 to 27 minutes: AI code review. Ask for the suggestion they rejected. Specific, slightly embarrassed stories score highest.
-
27 to 30 minutes: written communication. Give the time-zone scenario, then ask them to type the message they'd leave the designer into the chat. Two minutes of writing tells you more than ten minutes of talking.
Then pass the shortlist to an engineer for the technical test. Their hour goes on people who have already cleared the part that's hard to fake.
Anyone can generate a React Native app in 2027, so hire the person who knows which version to ship.














