Turkish Data Labeling & AI Safety Specialist

Welo Data; San Francisco (CA); On-site; USD 70,000 - 71,000; Full time

Benefits offered by this job

Gourmet dining

Snack suites

Commuter support

Full security benefits

Job summary

Welo Data is seeking a Data Labeling Associate for Project Perseus in San Francisco, CA. This position involves critiquing Arabic AI outputs for accuracy, identifying biases in datasets, and ensuring model improvement. Candidates must have native Turkish proficiency, a University degree, and 2+ years of experience in AI safety. The role offers campus benefits such as gourmet dining and comprehensive health coverage, fostering a collaborative environment in AI evaluation and improvement.

Qualifications

• 2+ years of experience focusing on Safety & Trust in Generative AI data delivery.
• 1–2 years of experience in professional writing, journalism, or content creation.
• Ability to make sound judgment calls in ambiguous scenarios.

Responsibilities

• Critique Arabic AI outputs for accuracy, fluency, and tone.
• Identify bias, risks, and failure modes in Arabic datasets.
• Apply complex linguistic guidelines to refine evaluation frameworks.
• Identify trends in model performance and communicate findings.
• Participate in team syncs to ensure consistent quality.

Skills

Native-level Turkish proficiency

B2 level English proficiency

Professional writing experience

Generative AI safety expertise

Critical thinking

Technical agility with LLMs

Education

University Degree (Bachelor’s or higher)

Job description

Welo Data is seeking a Data Labeling Associate for Project Perseus in San Francisco, CA. This position involves critiquing Arabic AI outputs for accuracy, identifying biases in datasets, and ensuring model improvement. Candidates must have native Turkish proficiency, a University degree, and 2+ years of experience in AI safety. The role offers campus benefits such as gourmet dining and comprehensive health coverage, fostering a collaborative environment in AI evaluation and improvement.