Data Scientist (NLP/ GENAI Evaluation)

Responsibilities
Collaborate with policy and communications officers, product owners, engineers, domain experts and subject-matter specialists to understand user requirements and translate them into well-defined analytical and machine-learning problems.
Build and maintain robust evaluation frameworks and datasets for natural-language, generative AI and other machine-learning applications.
Plan and carry out experiments to evaluate and enhance model quality, accuracy, consistency, reliability, latency and cost.
Assess suitable models, methods and emerging technologies, and recommend solutions based on evidence, user requirements and operational factors.
Conduct systematic error analysis, identify performance gaps across different use cases and user segments, and work with the team to prioritise areas for improvement.
Develop appropriate automated and human-evaluation methods, while recognising the limitations and risks associated with individual metrics and AI-assisted evaluation.
Work closely with engineers to implement validated improvements, establish quality checks and monitor performance in production environments.
Ensure that data, experiments and model-related decisions are reproducible, properly documented and aligned with responsible AI, privacy and security requirements.
Present findings, trade-offs and recommendations clearly to both technical and non-technical stakeholders.

Requirements:
A degree in Computer Science, Data Science, Statistics, Artificial Intelligence, Computational Linguistics or a related quantitative field, or equivalent practical experience.
Proven experience applying data science or machine learning to real-world problems, preferably in natural language processing, generative AI, search or information retrieval.
Strong Python programming skills, together with working knowledge of SQL, data processing, version control and software-development practices.
A solid understanding of statistics, experimental design, evaluation methods, sampling, error analysis and model validation.
Experience working with unstructured text or other complex data types, as well as evaluating machine-learning or generative AI systems using more than a single aggregate metric.
Familiarity with current NLP and AI concepts, including embeddings, language models, prompt design and model evaluation.
The ability to develop maintainable code and collaborate with engineers to deploy data-science solutions into production.
Strong analytical, problem-solving and communication capabilities, with the ability to explain technical findings and trade-offs clearly to a range of audiences.

Preferred Skills

Experience in multilingual NLP, translation-quality evaluation, or working alongside linguists and language reviewers.
Experience with cloud-based AI services, vector search, MLOps, production monitoring or responsible AI practices.
Experience developing AI-enabled products within government, communications, regulated sectors or other high-assurance environments.
An understanding of Singapore’s public communications landscape and the requirements of public-sector stakeholders.