Data Scientist (Artificial Intelligence)

Location

Job Scope

Credit Risk Modelling and Analysis (Approx. 50%)

• Research and develop next-generation credit scoring models using experimental machine learning techniques and non-traditional data sources.
• Design and execute experiments to test new features and hypotheses regarding credit risk behaviours in underserved populations.
• Collaborate with risk management teams to translate complex data findings into actionable credit assessment strategies.
• Enhance model robustness through rigorous A/B testing, back-testing, and validation frameworks.
• Address ad hoc data analysis requests with a focus on uncovering causal relationships.

AI Experimentation and Application Development (Approx. 50%)

• Conduct experiments on Large Language Models (LLM) to build autonomous agents and intelligent workflows.
• Design rapid prototypes and proof-of-concepts (PoCs) to test the feasibility of new AI technologies.
• Partner with engineering teams to integrate successful experimental models into our scalable LaaS platform.
• Optimize LLM inference and RAG (Retrieval-Augmented Generation) pipelines for cost and latency.

Other Duties

• Perform other related duties and ad-hoc tasks as assigned by the management to support the department's goals.

Requirements

Technical Skills

• Malaysian citizen with a Bachelor’s or Master’s degree in Data Science, AI, Computer Science, Statistics, or a related field.
• 1 - 4 years of experience in data science, with a strong portfolio of research projects or experimental modelling.
• Advanced proficiency in Python and SQL.
• Experimental Mindset: Demonstrated ability to set up experiments, track metrics, and draw statistical conclusions.
• Plus: Familiarity with cloud platforms (AWS, GCP) and containerization (Docker).

Professional Competencies

• Intellectual Curiosity: A strong desire to dig deep into "why" a model behaves the way it does, not just "how" to deploy it.
• Communication: Ability to explain experimental results and technical trade-offs to non-technical stakeholders.
• Resilience: Comfortable with the ambiguity of research work where experiments may fail before they succeed.