As a University of Applied Learning, the Singapore Institute of Technology (SIT) works closely with industry in its research pursuits. This position is situated within the SIT x NVIDIA AI Centre (SNAIC).

This role is part of an industry innovation project with a large consumer goods company, where you will develop an evaluation framework for vision-language model (VLM) with applications in the personal care sector. The research focuses on fine-grained VLM capabilities such as spatial reasoning, temporal grounding, event tracking, and domain knowledge using a curated multimodal dataset.

Key Responsibilities

• Manage the research project together with the Principal Investigator (PI) and industry partner to ensure all project deliverables are met
• Design and implement evaluation frameworks and metrics for vision-language models
• Develop annotated video datasets and capability-tagged evaluation tasks
• Build end-to-end evaluation pipelines and failure mode analysis tools to analyze VLM performance across reasoning dimensions
• Prepare technical reports, publications, and industry-facing deliverables
• Mentor student assistants
• Candidates are to communicate with any internal or external parties to ensure project deliverables are met.
• Any other ad-hoc duties as assigned by Supervisor.

Requirements

• PhD in Computer Science or related field
• Expertise in computer vision and vision-language models
• Experience with ML evaluation metrics and benchmarking
• Proficiency in Python and deep learning frameworks (e.g., PyTorch)
• Interest in applied, industry-collaborative research