Scientist

Agency for Science, Technology and Research (A*STAR); Singapore; On-site; SGD 90,000 - 150,000; Full time

Job summary

The FEAT Centre at A*STAR is seeking a Scientist to develop AI-enabled approaches for fusion plasma diagnostics and data analysis. You will work with diverse experimental data from fusion diagnostics to create robust analysis workflows and AI models.

The role offers opportunities to publish findings, collaborate with plasma physicists, diagnostic scientists, and international partners, and to contribute across both fusion diagnostics and AI model development.

Qualifications

• PhD in Physics, Engineering, Computer Science, Machine Learning, Applied Mathematics, Computational Science, or a closely related field.
• Candidates without prior industry or postdoctoral experience are encouraged to apply.
• Strong background in at least one of: plasma or experimental physics, scientific data analysis, machine learning, or scientific computing.
• Strong numerical and data-analysis skills, with proficiency in Python, C/C++, or Julia.
• Experience working with experimental, simulation, or other complex scientific datasets.
• Strong written and spoken English communication skills, with the ability to work effectively in interdisciplinary teams.

Responsibilities

• Develop and apply computational methods for the analysis and interpretation of fusion plasma diagnostic data.
• Process, curate, and analyse experimental datasets from multiple diagnostics and fusion devices.
• Develop and evaluate machine-learning models for diagnostic analysis, reconstruction, classification, and forecasting.
• Explore new AI approaches for learning from heterogeneous scientific data and transferring models across fusion devices.
• Develop scalable and reproducible analysis and machine-learning workflows using GPU and high-performance computing resources.
• Collaborate with internal teams and international research partners, publish research findings, supervise junior researchers and students, and contribute to collaborative projects and proposals.

Skills

Python

C/C++

Julia

Machine learning

Data analysis

Scientific computing

Education

PhD (Physics/Engineering/CS/ML/Applied Mathematics/Computational Science)

Tools

PyTorch

JAX

TensorFlow

GPU

Job description

Position Summary

The FEAT Centre is seeking a Scientist to develop AI-enabled approaches for fusion plasma diagnostics and data analysis. The work spans two closely connected areas: fusion diagnostics and data analysis, and AI model development, with flexibility to focus on either area or work across both.

Who we are looking for

The FEAT Centre is seeking a Scientist to advance the use of experimental diagnostics and AI for fusion plasma research. You will work with diverse experimental data from fusion plasma diagnostics, including microwave, imaging, and other measurement techniques, to develop new approaches for plasma analysis and interpretation. The research presents opportunities to develop AI methods that learn across heterogeneous measurements and experimental devices while incorporating knowledge of the underlying physics. This position is part of the Plasma Physics and Diagnostics Department (PPDD), which hosts expertise in plasma modelling and diagnostics. You will work closely with plasma physicists, diagnostic scientists, and AI researchers, and collaborate with international partners across universities and fusion research facilities. The research covers two complementary technical tracks: 1. Fusion diagnostics and data analysis: Analyse and interpret experimental diagnostic data, including signal processing, data curation, diagnostic reconstruction, validation, and the development of robust analysis workflows. 2. AI model development: Develop machine-learning methods for fusion diagnostic data, including multimodal and time-series modelling, representation learning, forecasting, transfer learning, and physics-informed approaches. Depending on your expertise, you may contribute primarily to one track or work across both. You will publish research findings, contribute to collaborative research projects, and help develop FEAT's capabilities in AI-enabled plasma diagnostics.

What you’ll be doing

• Develop and apply computational methods for the analysis and interpretation of fusion plasma diagnostic data.
• Process, curate, and analyse experimental datasets from multiple diagnostics and fusion devices.
• Develop and evaluate machine-learning models for diagnostic analysis, reconstruction, classification, and forecasting.
• Explore new AI approaches for learning from heterogeneous scientific data and transferring models across fusion devices.
• Develop scalable and reproducible analysis and machine-learning workflows using GPU and high-performance computing resources.
• Collaborate with internal teams and international research partners, publish research findings, supervise junior researchers and students, and contribute to collaborative projects and proposals.

What you’ll bring

Must-Haves

• PhD in Physics, Engineering, Computer Science, Machine Learning, Applied Mathematics, Computational Science, or a closely related field.
• Candidates without prior industry or postdoctoral experience are encouraged to apply.
• Strong background in at least one of the following areas: plasma or experimental physics, scientific data analysis, machine learning, or scientific computing.
• Strong numerical and data-analysis skills, with proficiency in scientific programming (e.g., Python, C/C++, Julia).
• Experience working with experimental, simulation, or other complex scientific datasets.
• Strong written and spoken English communication skills, with the ability to work effectively in interdisciplinary teams.

Good-to-Haves

• Background in plasma physics, fusion science, plasma diagnostics, or experimental physics.
• Experience with diagnostic or scientific data analysis, including signal processing, time-series analysis, imaging, or inverse problems.
• Experience developing machine-learning models using frameworks such as PyTorch, JAX, or TensorFlow.
• Experience with areas such as multimodal learning, representation learning, self-supervised learning, foundation models, forecasting, transfer learning, or uncertainty quantification.
• Experience with scientific or physics-informed machine learning.
• Experience with high-performance computing, GPU acceleration, or large-scale computational workflows.

What you’ll need to succeed

• Scientific problem-solving and critical thinking
• Independence, initiative, and willingness to learn
• Research communication (written and oral)
• Collaboration and teamwork in interdisciplinary environments

What we offer

• Opportunity to develop AI and diagnostic methods for challenging problems in fusion plasma research.
• Access to experimental datasets from international fusion research programmes and advanced GPU and high-performance computing resources.
• Close collaboration with plasma physicists, diagnostic scientists, AI researchers, and international fusion research partners.
• Flexibility to pursue research in fusion diagnostics and data analysis, AI model development, or across both areas.