About this role
United Overseas Bank seeks a GenAI Engineer/Scientist in its Innovation Group to improve and govern agentic applications across the bank. The role focuses on operating, evaluating and expanding agent-based solutions while maintaining secure, reliable and compliant services. This is a full-time, on-site position in Central Region (City Area), Singapore.
Responsibilities
- Improve agent performance and capabilities, and design, build and roll out solutions to additional users. Manage agent onboarding, configuration and retirement; monitor uptime, performance and service-level agreements.
- Work with business, technology and innovation teams to integrate agent workflows with incident, problem and change management. Maintain logging, audit trails and reporting for transparency and regulatory requirements. Investigate root causes, identify automation opportunities, optimize performance and produce SOPs, FAQs and user guides.
- Apply frameworks to evaluate agent performance, reliability and compliance. Lead triage and resolution of agent-related incidents and coordinate corrective actions.
- Develop and refine prompts, system messages and few-shot examples for relevance, accuracy and cost efficiency. Track latency, token usage and output quality; adjust prompts and pipelines to meet business and ethical requirements. Update roadmaps and prompts in response to regulatory, policy or business changes.
Qualifications
- Bachelor’s degree in Data Science, Statistics, Computer Science, Information Systems or a related field; 5 to 10 years of hands-on experience designing, building and operationalizing end-to-end AI/ML systems, including data preparation, model development, deployment and post-production monitoring. A minimum of 1 to 2 years must involve generative AI and large language model projects.
- Strong data science fundamentals and hands-on Python experience. Experience with LLM and agentic systems, fine-tuning, RAG, prompt optimization, evaluation, applied NLP and responsible AI practices. Experience designing prompts, tuning outputs and integrating LLMs such as Claude Sonnet 3.5 or GPT-class models into end-user applications; familiarity with AI agent evaluation techniques. Must communicate complex technical concepts to non-technical stakeholders.
- Banking or financial services experience is a plus.
Skills for this role
AI agentsGenerative AILarge language modelsPythonRetrieval-augmented generationPrompt optimizationFew-shot promptingModel evaluationNatural language processingEmbeddingsInformation retrievalClassificationInformation extractionModel fine-tuningExperimental designStatistical analysisError analysisPerformance monitoringService lifecycle managementProcess improvementIncident managementRoot cause analysisResponsible AIBias testingPrivacy leakage testingAudit trailsStakeholder communication