Arbetsbeskrivning

Veritaz is a leading IT staffing solutions provider in Sweden, committed to advancing individual careers and aiding employers in securing the perfect talent fit. With a proven track record of successful partnerships with top companies, we have rapidly grown our presence in the USA, Europe, and Sweden as a dependable and trusted resource within the IT industry.

Assignment Description:

We are looking for a Senior Traffic Analyst to join our dynamic team.

What you will work on:

• Develop and implement evaluation frameworks, scoring rubrics, and multidimensional assessment models to measure the quality of AI-generated and AI-translated content.
• Create, maintain, and refine datasets, scoring protocols, and language-specific guidelines to ensure accuracy, fluency, and cultural relevance.
• Collaborate with ML engineers and language experts to calibrate evaluators, analyze model errors, and improve assessment logic.
• Conduct prompt testing, model calibration, and scoring validation for Generative AI evaluation processes.
• Identify quality gaps, propose improvements, and contribute to continuous feedback loops that enhance the linguistic performance of AI platforms.
• Support the training of assessment models and contribute to documentation, process development, and adoption of best practices.

What you bring:

• Experience in language quality evaluation, applied linguistics, computational linguistics, or similar language-focused research environments.
• Hands-on experience with LLM evaluation, machine translation assessment, or annotation processes in multilingual contexts.
• Strong understanding of linguistic and translation assessment frameworks such as MQM, MetricX, or COMET, with experience creating scoring rubrics.
• Familiarity with Generative AI evaluation methods, including prompt testing, model calibration, and scoring validation.
• Experience collaborating with ML teams on data pipelines, annotation workflows, or fine-tuning of linguistic evaluators.
• Deep linguistic and cultural knowledge across multiple markets, with the ability to define and measure content quality for different contexts.
• Advantageous experience with programmatic QA using Python, YAML, gRPC, or rule-based validation.
• Advantageous knowledge of inter-rater reliability measurement methods such as Krippendorff’s alpha or Cohen’s kappa, and experience comparing human and AI-generated assessments
• Fluent in English and Swedish.