Arabic Language & Technical Consultant（Part-time）

Job Introduction ：

We are recruiting talents proficient in Arabic dialects of Saudi Arabia and the Gulf region to support our AI dialect model project for financial call centers.

You will conduct dialect research, provide business and data consulting, identify differences and application priorities of various dialects in financial customer service scenarios, and support data collection and model optimization.

1.

Research the application of Najdi, Hijazi, Gulf and other dialects in scenarios including financial customer service, outbound collection and customer complaint handling;

2.

Sort out differences in vocabulary, financial expressions and emotional expressions across dialects, and provide reference cases;

3.

Put forward suggestions on dialect selection and coverage priority for banking, insurance and other businesses;

4.

Plan channels for Arabic corpus collection, and assess the feasibility and compliance risks of data sources;

5. 【Optional】Participate in the evaluation of ASR and LLM models, and verify the localization adaptability of model outputs.

Job Requirements ：

1.

Saudi national or long-term resident in Saudi Arabia; proficient in either Najdi Arabic or Hijazi Arabic.

Familiarity with Gulf Arabic and White Arabic / Educated Spoken Arabic is required preferably, with a good understanding of local language and culture.

2.

Prior working experience in financial call centers, banking, insurance, debt collection or related fields is preferred.

3.

Must be bilingual in Arabic and English for daily communication.

4.

Able to work remotely on a part-time basis, attend weekly meetings on time and complete assigned research and documentation tasks.

Working experience in corpus annotation, AI model training, speech evaluation and other related fields is a plus.

-Remote work, task-based arrangement

-Hourly Rate: $20 per hour

-Project Info: Long-term project.

Outstanding performers will have opportunities to participate in follow-up work such as model iteration.