Postdoctoral Researcher in Fine-tuning Large Language Models and Agentic Systems for Physics and Materials Science (Forschungszentrum Julich Institute for Advanced Simulation)

Join our team until the next possible date as Postdoctoral Researcher – Fine-tuning Large Language Models and Agentic Systems for Physics and Materials Science The Institute for Advanced Simulation – Materials Data Science and Informatics (IAS-9) focuses on innovative computational methods, combining machine learning/AI with high-performance computing to address a broad range of scientific challenges. Institute issuing the offer: IAS-9 Your Job: ...to revolutionize applications of the physical sciences through tailored large language models (LLMs) and intelligent agentic systems. Your tasks in detail: Development and fine-tuning of AI models and LLMs specifically tailored for data mining in the physical sciences and engineering Fine-tuning and evaluation of open-source LLMs (eg, Teuken 7B, Llama 3, Mistral 7B) specifically for physics and materials science applications scenarios as well as for related research data management tasks Developing LLM-driven agents that integrate domain knowledge to support structured chain-of-thought processes Benchmarking and validating models against domain-specific tasks such as retrieval, summarization, metadata extraction, and structured problem-solving Investigating advanced strategies for model adaptation, including in-domain pre-training, domain adapters, task-specific instruction tuning, and retrieval-augmented generation (RAG) Collaboration with interdisciplinary teams to enhance model performance, adaptability, and usability within physics and materials science communities Publication and dissemination of research findings in scientific journals and conferences Your Profile: A masterdegree and a doctoral degree in Physics, Materials Science, Computer Science, Data Science, or related fields Proven experience with large language models (LLMs), natural language processing (NLP), and fine-tuning techniques Familiarity with structured reasoning, chain-of-thought processes, and agent-based systems is beneficial Strong programming skills (preferably Python); experience with high-performance computing (HPC) infrastructures is advantageous Excellent analytical and problem-solving capabilities Proven track record of publications in reputable scientific journals and at the leading conferences Ability to work effectively in interdisciplinary and international teams Good communication and writing skills in English Our Offer: We work on current challenges that impact our society and are offering you the chance to actively help in shaping the change! We support you in your work with: Secure employment in a dynamic, international and multidisciplinary environment of scientists from different domains and software developers A creative environment, with an open-minded team (local and across Germany) with heterogeneous backgrounds and expertise Extensive occupational health management and a variety of sports activities (eg beach volleyball court, running groups, yoga classes, and much more) Flexible working hours in a full-time position (39 hours/week) with the option of slightly reduced working hours Flexible work arrangements, eg working from home 30 days of annual leave and provision for days off between public holidays and weekends (eg between Christmas and New Year) Capital-forming benefits and an employee pension scheme In addition to exciting tasks and a collaborative working atmosphere at Jülich, we have a lot more to offer: https://go.fzj.de/benefits Place of employment: Jülich/Aachen You will be initially employed for a fixed term of 2 years with the prospect of longer-term employment. Salary and social security benefits will conform to the provisions of the Collective Agreement for the Public Service (TVöD-Bund), pay group 13, depending on your current qualifications and the precise nature of the tasks assigned to you. We welcome applications from people with diverse backgrounds, including in terms of age, gender, disability, sexual orientation/identity, and social, ethnic and religious origin. A diverse and inclusive working environment with equal opportunities in which everyone can realize their potential is important to us. We are looking forward to your application including a motivation letter, CV, and relevant certificates. Relevant experience must be proven and be briefly highlighted in the mandatory letter of motivation. We are looking to fill this position as soon as possible – the job advertisement will be taken offline once the position is filled. Further information on diversity and equal opportunities: https://go.fzj.de/equality The job will be advertised until the position has been successfully filled. You should therefore submit your application as soon as possible. We look forward to receiving your application via our Online-Recruitment-System! Questions about the vacancy? Get in touch with us by using our contact form Please note that for technical reasons we cannot accept applications via email. Pdf If you apply for this position please say you saw it on Engineeroxy

Join our team until the next possible date as

Postdoctoral Researcher – Fine-tuning Large Language Models and Agentic Systems for Physics and Materials Science

The Institute for Advanced Simulation – Materials Data Science and Informatics (IAS-9) focuses on innovative computational methods, combining machine learning/AI with high-performance computing to address a broad range of scientific challenges.

Institute issuing the offer: IAS-9

Your Job:

...to revolutionize applications of the physical sciences through tailored large language models (LLMs) and intelligent agentic systems.

Your tasks in detail:

Development and fine-tuning of AI models and LLMs specifically tailored for data mining in the physical sciences and engineering
Fine-tuning and evaluation of open-source LLMs (eg, Teuken 7B, Llama 3, Mistral 7B) specifically for physics and materials science applications scenarios as well as for related research data management tasks
Developing LLM-driven agents that integrate domain knowledge to support structured chain-of-thought processes
Benchmarking and validating models against domain-specific tasks such as retrieval, summarization, metadata extraction, and structured problem-solving
Investigating advanced strategies for model adaptation, including in-domain pre-training, domain adapters, task-specific instruction tuning, and retrieval-augmented generation (RAG)
Collaboration with interdisciplinary teams to enhance model performance, adaptability, and usability within physics and materials science communities
Publication and dissemination of research findings in scientific journals and conferences

Your Profile:

A masterdegree and a doctoral degree in Physics, Materials Science, Computer Science, Data Science, or related fields
Proven experience with large language models (LLMs), natural language processing (NLP), and fine-tuning techniques
Familiarity with structured reasoning, chain-of-thought processes, and agent-based systems is beneficial
Strong programming skills (preferably Python); experience with high-performance computing (HPC) infrastructures is advantageous
Excellent analytical and problem-solving capabilities
Proven track record of publications in reputable scientific journals and at the leading conferences
Ability to work effectively in interdisciplinary and international teams
Good communication and writing skills in English

Our Offer:

We work on current challenges that impact our society and are offering you the chance to actively help in shaping the change! We support you in your work with:

Secure employment in a dynamic, international and multidisciplinary environment of scientists from different domains and software developers
A creative environment, with an open-minded team (local and across Germany) with heterogeneous backgrounds and expertise
Extensive occupational health management and a variety of sports activities (eg beach volleyball court, running groups, yoga classes, and much more)
Flexible working hours in a full-time position (39 hours/week) with the option of slightly reduced working hours
Flexible work arrangements, eg working from home
30 days of annual leave and provision for days off between public holidays and weekends (eg between Christmas and New Year)
Capital-forming benefits and an employee pension scheme

In addition to exciting tasks and a collaborative working atmosphere at Jülich, we have a lot more to offer: https://go.fzj.de/benefits

Place of employment: Jülich/Aachen

You will be initially employed for a fixed term of 2 years with the prospect of longer-term employment. Salary and social security benefits will conform to the provisions of the Collective Agreement for the Public Service (TVöD-Bund), pay group 13, depending on your current qualifications and the precise nature of the tasks assigned to you.

We welcome applications from people with diverse backgrounds, including in terms of age, gender, disability, sexual orientation/identity, and social, ethnic and religious origin. A diverse and inclusive working environment with equal opportunities in which everyone can realize their potential is important to us. We are looking forward to your application including a motivation letter, CV, and relevant certificates. Relevant experience must be proven and be briefly highlighted in the mandatory letter of motivation. We are looking to fill this position as soon as possible – the job advertisement will be taken offline once the position is filled.

Further information on diversity and equal opportunities: https://go.fzj.de/equality

The job will be advertised until the position has been successfully filled. You should therefore submit your application as soon as possible. We look forward to receiving your application via our
Online-Recruitment-System!

Questions about the vacancy?
Get in touch with us by using our contact form
Please note that for technical reasons we cannot accept applications via email.

Pdf

If you apply for this position please say you saw it on Engineeroxy

Apply

All Jobs