The rarest role in AI — and the most in-demand in the Arab world

1. TII (Abu Dhabi) — the Falcon family. Pre-training, post-training, multi-modal research. The most prestigious AI research home in the region. Hiring across senior researcher and applied research scientist levels.
2. G42 Inception — applied Arabic LLM teams, evaluation and red-teaming, partnerships with international labs. More applied than TII; less pure research.
3. SDAIA / NCAI — the National Centre for AI in Riyadh. Government-aligned, mandate to build Arabic-first AI capability for the kingdom. ALLaM and related work.
4. MBZUAI — graduate research university in Abu Dhabi. Faculty positions and research scientist roles, particularly inside the NLP and ML departments. Open to genuine career academics.
5. HUMAIN — the new PIF-backed entity in KSA. Building an Arabic LLM programme alongside a broader AI agenda. Smaller, earlier, but moving fast.
6. Lucidya (KSA) — applied Arabic NLP for customer experience and social listening. The most established commercial Arabic NLP team in the region.

There are three sources:

1. Diaspora researchers in UK and US universities — UCL, Cambridge, MIT, Stanford, CMU. Typically open to a relocation conversation if the project is right and the brief is honest.
2. MBZUAI graduates — the first cohorts are now 2–4 years post-PhD. Strong technical depth, often a natural fit for TII or G42 roles.
3. International senior researchers returning from labs — DeepMind, Meta AI, Anthropic, OpenAI. Smaller in number, but increasingly visible. Several high-profile moves in the last 18 months.

The size of the global pool that can credibly do this work — publish in Arabic NLP venues, fine-tune Arabic LLMs, or run evaluation programmes on Arabic corpora — is probably under 500 people worldwide. The market is small enough that most senior researchers know each other by name.

"Less than 0.5% of global NLP research is focused on Arabic — a structural gap the best-capitalised AI organisations are all trying to close at the same time."

Scarcity. The premium applies on top of an already-elevated ML engineering benchmark, so total packages at the senior level cluster at the top of the engineering band. In practice the premium tends to be expressed as a higher bonus, a larger relocation budget, or a research stipend, rather than a higher base — because base salary is more visible and benchmarked internally, while bonus and ancillary structures are more flexible.

For the most senior researchers, structured equity-equivalent LTIP packages are now appearing at TII and G42 — typically vesting over 3 years against research milestones.

Clients look for one or more of:

• Published research at ACL, EMNLP, NAACL, or similar venues, on Arabic-specific work
• Hands-on experience with Arabic pre-training or continual pre-training
• Evaluation methodology — building or using Arabic-specific benchmark suites
• Speech recognition or text-to-speech work on Arabic dialectal variants
• Industry experience deploying Arabic LLM systems in production

A MENA background is not required. Strong Arabic literacy is. A small number of candidates have moved into senior Arabic NLP roles with limited Arabic fluency, but only at the most senior leadership levels — where the work is direction-setting rather than hands-on linguistic engineering.