Language AI Solutions Specialist

As a Language AI Solutions Specialist, you will bridge the gap between traditional localization and cutting-edge AI. You will be responsible for the end-to-end lifecycle of translation systems,from legacy Neural MT to modern LLM-based solutions. Your mission is to architect, tune, and deploy high-performance language models that drive global scalability while maintaining uncompromising linguistic quality.

Key Deliverables

• Hybrid Engine Deployment: Design, test, and certify Neural MT systems (KantanMT) and fine-tuned LLMs tailored to client-specific domains.
• Performance Analytics: Deliver comprehensive quality reports, including Predictive TER scores, MQM-DQF/COMET/G-Eval metrics, and Human-in-the-Loop (HITL) evaluations.
• Production Integration: Working closely with internal stakeholders (Development, Production, etc.) to oversee the seamless transition of validated models into live production environments.

Core Responsibilities

AI Model Tuning & Optimization

• Fine-Tuning & Hyperparameters: Execute supervised fine-tuning (SFT) on LLMs to align output with Client/Project specific style, voice, and domain-specific terminology.
• RAG Corpus Management: Architect and maintain high-quality corpora for Retrieval-Augmented Generation, ensuring that vector databases are populated with clean, relevant, and deduplicated data.
• Data Engineering: Perform advanced data analysis, cleansing, and data generation to prepare clean datasets for model training and tuning.
• Output Verification: Utilize native-level linguistic expertise to perform deep-dive audits of NMT and LLM outputs, identifying nuanced errors such as hallucinations, cultural insensitivity, or tone inconsistency that automated metrics might miss.

Linguistic Asset Engineering

• Modernization: Manage and optimize Translation Memories (TMs) and Termbases to serve as the foundational "ground truth" for AI training.
• Quality Control: Implement and evolve Machine Translation Post-Editing (MTPE) workflows, integrating AI-assisted quality estimation (QE) to reduce manual overhead.

Analytics & Productivity Insights

• Impact Metrics: Gather and analyze post-deployment data, focusing on Edit Distance, time-to-market reduction, and cost-per-word efficiency.
• Stakeholder Reporting: Translate complex technical data into actionable insights for Program Managers and clients, recommending rebuilds or retraining cycles where necessary.

Operations & Troubleshooting

• Technical Support: Serve as the escalation point for engine failures, API latency issues, or unexpected model "hallucinations".
• Process Documentation: Create "Living Documentation" for evolving AI workflows to ensure team-wide alignment and scalability.

Requirements

Required Skills

• Fluent English and native level of any core Asian languages (Chinese, Japanese or Korean)
• 3-4 years of experience in the localization industry, working as a Project Manager, Translator, Post-Editor or QA tester
• Machine Translation Post Editing: Solid understanding of MTPE, proven track record in correcting NMT or LLM outputs for fluent, accurate and brand-consistent target text.
• Degree in Translation, Translation Technology or relevant experience.
• Aptitude for analytics: Ability to understand and work with string-matching and edit metrics (e.g., explaining Edit Distance in a post editing context, understanding MQM-DQF frameworks etc.)

Preferred Skills

• Solid understanding of CAT tools and CMS/TMS systems in the localization industry
• AI/LLM applications - knowledge of prompt engineering, RAG corpus management and familiarity with Vector Spaces
• Basic knowledge of Python and Regex