Senior Manager / Manager, Large Language Model (LLM) Architect – SIMFONI

About CRIS

The Consortium for Clinical Research and Innovation Singapore (CRIS), fully owned by MOH Holdings, aims to lead transformative strategies for national clinical research and development initiatives under the Singapore Ministry of Health. CRIS unites seven national R&D and service programmes, including SIMFONI, with the goal of advancing research capabilities for a robust and future-ready healthcare system.

About SIMFONI

The Singapore Medical Foundation AI Model (SIMFONI) Programme was launched in 2025 to foster the responsible use of AI in public healthcare, supporting healthcare professionals in delivering optimal patient care. The programme is focused on building the infrastructure and expertise needed for developing and deploying large-scale foundation models to improve healthcare outcomes.

Position Overview

The organisation seeks a senior technical architect to steer the direction of a healthcare foundation model initiative centred around large language models (LLMs). This leadership position involves overseeing the adaptation, development, evaluation, and integration of LLMs for medical use cases such as clinical decision support and knowledge management. The successful applicant will collaborate cross-functionally, lending expertise in proposal review, technical guidance, risk assessment, and feasibility analysis for multiple technical teams.

Key Responsibilities

• Serve as the technical lead on LLM architecture, guiding strategies for model selection, training, tuning, evaluation, and engineering feasibility.
• Evaluate and advise on technical proposals regarding continual pretraining, fine-tuning, evaluation design, infrastructure, and deployment risk.
• Assess proposed methods for their technical soundness, scalability, feasibility, and alignment with project aims.
• Establish and disseminate best practices, reference methodologies, and engineering guidelines throughout the programme.
• Advise on advanced LLM topics such as domain-adaptive pretraining, supervised/instruction tuning, preference optimization, retrieval-augmented generation (RAG), and safety/robustness measures.
• Support review of experiment designs, training stability, reproducibility, versioning strategies, and clinical safety evaluations.
• Guide engineering teams on model development workflows, evaluation pipelines, lifecycle management, observability, and deployment management.
• Collaborate with platform, infrastructure, data, and solution teams to ensure cross-project technical alignment and feasibility.
• Identify and communicate engineering risks associated with scalability, reliability, maintainability, costs, and operations.
• Coordinate technical alignment sessions, facilitate design reviews, and manage governance and documentation processes.
• Translate clinical and technical needs into actionable criteria and recommendations for project teams.

Requirements

• At least 5 years in AI/LLM, machine learning, or large-scale tech architecture roles.
• Hands-on experience in developing, adapting, evaluating, or deploying LLM/foundation models.
• Proficiency in modern LLM architectures, training and tuning methods, model evaluation, and benchmarking practices.
• Strong background in continual or domain-adaptive pretraining, supervised and instruction tuning, RLHF/DPO/GRPO, RAG, safety testing, and hallucination mitigation.
• Knowledge of distributed training, GPU infrastructure, model serving, and MLOps/LLMOps concepts.
• Familiarity with model management, reproducibility, observability, and production deployment standards.
• Proven ability to critically assess technical designs, challenge assumptions, and provide actionable advice.
• Demonstrated engineering judgment focusing on model quality, scalability, reliability, maintainability, and cost.
• Prior experience in technical leadership or advisory roles across multiple teams or organisations.
• Excellent communication skills, including the ability to explain complex technical concepts to diverse audiences and produce thorough documentation.

Preferred

• Experience in production or near-production deployments of LLMs or foundation models.
• Background in healthcare AI, biomedical NLP, or medical data platforms.
• Understanding of data governance, clinical safety, privacy, and responsible AI practices in a healthcare context.
• Involvement in multi-entity, partner-driven technology programmes.
• Awareness of Singapore’s public healthcare landscape and national healthcare technology initiatives.

Additional Information

• This role is a 3-year renewable contract based in Singapore and only shortlisted candidates will be contacted.
• CRIS does not request payments or sensitive personal/financial information via unauthorised channels. Verify opportunities through official CRIS communications.