Research Fellow (AI & Large Language Models for Mental Health)

Requirements

• PhD in computer science, machine learning, data science, computational linguistics or social science, or a closely related field.
• Strong expertise in deep learning frameworks (PyTorch, JAX, or TensorFlow); hands-on experience with LLMs or foundation models is a significant advantage.
• Proficiency in Python; familiarity with R is beneficial but not required.
• Experience in at least two of: LLM fine-tuning and alignment (SFT, RLHF/RL), NLP, multi-agent systems, reinforcement learning, affective computing, dialogue systems, or privacy-preserving machine learning.
• Demonstrated publication record commensurate with career stage; genuine interest in mental health and real-world public-health impact, with the ability to communicate technical work to non-specialist audiences.
• Desirable: prior work in digital health or clinical NLP, experience handling sensitive or regulated data, and cross-cultural / multilingual (Singapore-context) sensitivity.
• Application Process
• The group supports career development through conference travel, training workshops, and mentorship from senior researchers and international collaborators. Interested applicants should submit:
• A comprehensive curriculum vitae, including a full list of publications.
• A research statement (maximum two pages) outlining past contributions and future directions.
• Contact information for two professional references (letters may be requested).
• Review of applications begins immediately and continues until the position is filled; the anticipated start date is flexible by negotiation.
• NUS is an equal opportunity employer committed to diversity and inclusion. We welcome applications from all qualified individuals regardless of race, gender, religion, national origin, age, or disability.
• Enquiries: contact Dr. He Kai at Kai_he[at]nus[dot]edu[dot]sg with the subject line "RF Application - AI & LLMs for Mental Health".

Additional Information

Interested applicants are invited to apply directly at the NUS Career Portal. Please note your application will only be processed if you apply via NUS Career Portal. NUS Career Portal link: https://careers.nus.edu.sg/job/Research-Fellow-%28AI-&-Large-Language-Models-for-Mental-Health%29/33331-en_GB/?st=012EF294E13BD9D38C4C70B9339CF8E158925A52 We regret that only shortlisted candidates will be notified. Job Description Institution: NUS Saw Swee Hock School of Public Health, Singapore (SSHSPH) Level: Research Fellow Salary: Commensurate with experience and qualifications Duration: 1 year, renewable subject to performance and funding Start date: Flexible by negotiation (target: from 08/2026) The Saw Swee Hock School of Public Health (SSHSPH) at the National University of Singapore invites applications for a Research Fellow to drive AI-centred research on empathetic Large Language Models (LLMs) for mental health. The project builds a closed-loop system integrating low-cost screening at scale, stepped-care triage, intervention, and continuous evaluation - expanding access to mental-health support at population level while preserving clinician oversight for higher-risk cases. At its core is a triad of LLM agents (Counsellor, Client, and Evaluator) that co-evolve through iterative self-play across an assessment → diagnosis → intervention → re-evaluation workflow, with emphasis on Singapore-context localisation, ethical deployment, and privacy-by-design. The postholder will work with Dr. He Kai (Principal Investigator, SSHSPH) alongside collaborators across NUS, Nanyang Technological University (NTU), and the MOH Office for Healthcare Transformation (MOHT), which operates Singapore's national digital mental-health platform, mindline.sg. They will have significant latitude to shape their research direction and will collaborate closely with engineers, clinicians, psychologists, and policy stakeholders. Develop and validate the closed-loop Counsellor-Client-Evaluator triad of LLM agents for assessment, diagnosis, intervention planning, and re-evaluation. Design and implement supervised fine-tuning (SFT) and reinforcement learning (RL) pipelines for empathetic, clinically appropriate, and culturally congruent responses. Build rubric-driven evaluation across cognitive, emotional, behavioural, physiological, and social domains; develop automated reporting with standardised scales (e.g., PHQ-9, GAD-7) and RAG-based localisation. Apply privacy-preserving techniques (federated learning, differential privacy, data minimisation) and support deployment on mindline.sg with ethical safeguards and escalation protocols for high-risk cases. Publish findings in leading venues, co-supervise junior researchers, and contribute to grant reporting and proposal writing.

Your Match

How well this role fits your profile.

Company Intel

What employees say

Worked at NATIONAL UNIVERSITY OF SINGAPORE? Share your experience