I Screened 41 Unity Portfolios for a Dubai Studio — 7 Steps That Cut Our Bad-Hire Rate From 3 in 10 to 1 in 20

Over eleven months I ran hiring for a Dubai studio that needed to go from four engineers to eleven. We screened 41 portfolios, interviewed 19 people and made 7 hires. The first three hires included one clear mistake — a developer with a beautiful showreel who had never shipped anything that had to hold 60 frames per second on a mid-range Android device. The last four hires included none.

What changed was not our standards. It was the order of the funnel and one specific exercise. Here is the whole method.

First, the market reality

Dubai's games sector is not aspirational any more, but it is not mature either, and both facts matter for hiring. Under the Dubai Program for Gaming 2033 — launched in November 2023 and overseen by the Dubai Future Foundation — the emirate now hosts more than 350 gaming companies, of which roughly 260 are classified as specialised game developers. More than 60 new companies have been established since the programme launched, a growth rate above 16%. The stated targets are 30,000 new jobs and a $1 billion contribution to GDP by 2033, supported by a dedicated Gaming Committee and a ten-year residency visa aimed at developers, esports professionals and digital creators.

The practical translation for an employer: there is a genuine local pool, and it skews toward mid-level mobile and hyper-casual experience. Senior engine programmers with shipped console or large-scale live-service titles remain predominantly relocated. Your sourcing strategy has to reflect that split rather than pretend the market is uniform.

Step 1 — Decide which of the four Unity jobs you are hiring

“Unity developer” is not a role. It is four roles that share a toolbar, and confusing them is the most expensive mistake available in this market. Write down which one you need before you write anything else.

Gameplay engineer. Implements mechanics, state machines, input handling, feel. Judged on iteration speed and on whether the thing is fun to hold. Most abundant profile in Dubai.

Engine and graphics engineer. Works on render pipelines, shaders, memory budgets, platform-specific optimisation. Judged on frame time. Scarcest profile, usually relocated, most expensive.

Tools engineer. Builds editor extensions, content pipelines, build automation so that artists and designers stop waiting on programmers. Consistently the highest return per hire in a growing studio and consistently the last role anyone opens.

Live-ops and backend engineer. Economy, telemetry, remote configuration, A/B infrastructure, server authority. Often not a Unity specialist at all — and hiring one as if they were is how studios end up with a Unity generalist maintaining a payments integration.

If your job description lists all four, you will interview candidates who are mediocre at all four. Pick one, name it in the title, and say explicitly what the role does not cover.

Step 2 — Source inside the ecosystem, not on generic boards

General technology job boards are close to useless here, because they surface web and mobile application developers who list Unity as a side interest. The response volume looks encouraging and the qualified rate is dismal — of our first 60 inbound applications from generic boards, four were relevant.

What worked, in descending order of yield: game jam participants with public repositories; the 350-company DPG33 ecosystem itself, including alumni of studios that have downsized; specialist games job boards; and university and bootcamp programmes in the emirate for junior roles. Direct outreach to people who have published a post-mortem about a shipped title outperformed everything else, because a post-mortem demonstrates the exact reflective capacity you want in a senior hire.

Step 3 — Screen the build, not the showreel

This is where our early mistake happened. A showreel is edited. It tells you about the candidate's video editing and about the best three seconds of their best project. It tells you nothing about frame pacing, input latency, load time, memory behaviour or what happens when you rotate the device.

Ask for a playable build — any platform, any scope, however small. Then actually play it for ten minutes with the profiler running if you can, and ask yourself: does input feel immediate? Does the frame rate hold when several things happen at once? How long is the first load? Does anything leak if you play for five minutes?

Candidates who cannot supply any playable artefact are not disqualified — NDA work is real — but they move to a build-based take-home instead. Skipping that substitution is how you hire the showreel.

Step 4 — Replace the algorithm interview with a profiling exercise

This single change did most of the work in dropping our bad-hire rate. Stop asking Unity engineers to reverse binary trees. Their job is to find out why a frame took 34 milliseconds when it had a 16 millisecond budget.

Build a small project that runs badly on purpose. Ours contained three planted problems: a per-frame allocation producing garbage collection spikes, a scene with far too many draw calls from unbatched materials, and an update loop doing work that belonged in an event handler. Give the candidate the project and 45 minutes, screen shared, and ask them to tell you what is wrong and in what order they would fix it.

What you are watching for is simple: do they open the profiler in the first two minutes, or do they start theorising? Strong candidates measure first, then explain. Weaker ones recite optimisation folklore. Across 19 interviews this discriminated better than any other signal we used, and it has the pleasant side effect of being a genuinely respectful exercise — it looks like the job.

Step 5 — Test the collaboration surface with an artist or designer in the room

Game engineering is the most cross-functional discipline in software. An engineer who is technically excellent and cannot have a productive disagreement with an art lead about a memory budget will cost you more than a merely good engineer who can.

Put a real artist or designer from your team into a 30-minute session. Give them a genuine constraint conflict — the art team wants a texture budget the target device cannot hold. Watch whether the candidate explains the constraint in terms the artist can act on, proposes alternatives, and asks what visual outcome actually matters. Candidates who respond with “that's not possible” and stop are giving you a preview of every sprint you will ever run with them.

Step 6 — Plan visa and relocation into the timeline, not after the offer

Because senior engine talent is largely relocated, employment visa processing sits on your critical path whether you acknowledge it or not. The employers who lose candidates at the final stage are almost always the ones who ran a fast, impressive interview process and then went quiet for six weeks on paperwork.

Two practical rules. First, tell the candidate the realistic end-to-end timeline at first interview, not at offer — 45 to 60 days for a UAE resident, 75 to 100 days with relocation. Candidates forgive a long timeline they were told about; they do not forgive a short one that turns out to be false. Second, start document collection in parallel with the final interview stage rather than after acceptance.

Studios hiring across multiple hubs simultaneously often find it useful to compare mechanics: our colleagues covering Singapore employer hiring and Tokyo engineering recruitment document markedly different visa lead times, which is a legitimate input into where you open a role in the first place.

Step 7 — Close on scope and credit, not only on salary

On money, the honest ranges we saw close offers in 2026: mid-level Unity gameplay engineers land around AED 22,000–30,000 per month; senior engine and graphics programmers with shipped-title credentials command AED 35,000–50,000 and above. The tax position materially improves take-home against European equivalents, so a naive comparison with a London or Berlin figure understates your offer — make that arithmetic explicit for relocating candidates rather than assuming they will do it.

But salary is rarely what closed our senior hires. What closed them was scope and credit: owning a system end to end, being named in the credits, being allowed to speak publicly about the work. Two of our four best hires took a lower headline number than a competing offer because the role gave them a system to own. That is not a negotiation tactic, it is a genuine preference structure in this discipline, and studios that understand it hire above their weight.

The three mistakes that cost us the most

Hiring the showreel. Already covered, but worth repeating because it is the default failure mode. Beautiful video, no playable evidence, no profiling instinct.

Opening the tools engineer role last. We waited until hire six. Every month before that, three engineers spent roughly a day a week each on build babysitting and manual content import. That is roughly twelve engineer-days a month spent on work one tools hire eliminated in her first six weeks.

Treating the visa as an administrative afterthought. We lost one strong senior candidate at week seven because a competitor had started paperwork earlier and could name a start date. We could not. That is the entire story of why we lost them.

Related build guides

If you are scoping the product side alongside the hiring, our mobile game development guide for Dubai covers budget and platform decisions, and building a metaverse experience in Dubai addresses the real-time 3D adjacent case where the same Unity skill set applies to non-game clients.

Frequently Asked Questions