Research Assistant / Machine Vision Engineer

NATIONAL UNIVERSITY OF SINGAPORE; Singapore; On-site; SGD 20,000 - 60,000; Full time

Job summary

A leading university in Singapore is seeking a Research Assistant to support machine vision research projects focused on vision-language model development. The role involves finetuning models for image analysis tasks, implementing multimodal systems, and evaluating model performance. Candidates should have a degree in Computer Science and be experienced in writing well-documented code. Strong collaboration skills are essential, while experience with large language models is desirable. This position is ideal for individuals passionate about advancing technology in the cultural heritage domain.

Qualifications

• A degree in Computer Science or equivalent qualification.
• Demonstrable experience writing well-documented code is essential.

Responsibilities

• Finetune vision-language models for image segmentation and classification applications.
• Implement and maintain multimodal retrieval-augmented generation systems.
• Conduct experiments to evaluate model performance and document methodologies.

Skills

Experience writing well-documented code

Collaboration skills

Knowledge of transformer architectures

Education

Degree in Computer Science or equivalent qualification

Job description

Interested applicants are invited to apply directly at the NUSCareer Portal

Your application will be processed only if you apply via NUSCareer Portal

We regret that only shortlisted candidates will be notified.

Job Description

Role Overview

The Research Assistant will support machine vision research projects in the cultural heritage domain, focusing on the development and optimization of vision-language models for image analysis tasks.

Primary Responsibilities

• The position involves finetuning vision-language models (VLMs) for image segmentation and classification applications, as well as the systematic evaluation and benchmarking of model performance.
• The Research Assistant will implement and maintain multimodal retrieval-augmented generation (RAG) systems, conduct experiments to evaluate model performance, and document methodologies and results. The role requires collaboration with the research team to iterate on model architectures and training approaches based on experimental outcomes.

Required Qualifications

• A degree in Computer Science or equivalent qualification.
• Demonstrable experience writing well-documented code is essential. This includes clear code structure, comprehensive comments, and maintained documentation that facilitates collaboration and reproducibility.

Desirable Qualifications

• Experience training or finetuning large language models (LLMs) is desirable.
• Familiarity with transformer architectures and optimization techniques will be considered advantageous.
• Prior experience in the cultural heritage domain is not required.