About the role
KOPI RECRUIT PTE. LTD. is seeking a Data Scientist to support 4-5 product squads working with government agencies to redesign workflows and build AI-enabled solutions. You will work on problems involving large volumes of operational data and help teams move beyond rules-based automation.
The role covers exploratory analysis, modelling, building data pipelines, and deploying models into production while collaborating with domain experts to ensure analytics are integrated into real-world workflows.
Analyse large operational datasets to identify patterns, anomalies, outliers and risk indicators. Develop anomaly detection, clustering, segmentation and risk scoring approaches. Build supervised ML models where historical outcomes or labeled cases are available. Perform feature engineering using structured and unstructured data. Design evaluation approaches to assess model performance, including precision, recall, false positives and false negatives. Work with domain experts to determine how analytical outputs should be incorporated into operational workflows. Build and maintain data-processing and feature pipelines to support modelling at scale. Develop production-ready model or scoring pipelines and integrate them with applications built by forward-deployed teams. Monitor model performance, data quality and data drift after deployment. Conduct exploratory analysis to identify new opportunities for workflow redesign and automation. Support the use of LLMs and other AI techniques alongside traditional statistical and machine-learning approaches where appropriate. Develop reusable analytical approaches and components that can be applied across multiple agency engagements. Experience with large, messy operational or transactional datasets. Experience in applied data science, including statistics, feature engineering, machine learning and model evaluation. Experience with techniques such as anomaly detection, classification, clustering or risk modelling. Experience building data pipelines and taking models beyond experimentation into production environments.