Postdoctoral Scholar in Agentic ML Development for Rare Event Physics

Position Description

Scope of Research:

The Halıcıoğlu Data Science Institute (HDSI) at the University of California San Diego invites applications for a Postdoctoral Scholar in Agentic ML Development for Rare Event Physics.

The position will focus on developing next-generation agentic AI systems for scientific discovery, with a particular emphasis on searches for dark matter and neutrinos. The postdoctoral scholar will develop LLM-agent systems that engage deeply with physics-discovery workflows and accelerate key stages of complex experimental and data-analysis pipelines.

The position builds on the group’s recent development of SIDERIUS, an agentic framework for long-horizon machine-learning development. SIDERIUS has been demonstrated across multiple physics time-series datasets and, on the TIDMAD axion dark-matter benchmark, achieved the strongest valid denoising score among all agents benchmarked in the study.

The successful candidate will be expected to build beyond SIDERIUS rather than simply maintain it, using it as an existing platform while developing new methods and applications. This may include extending SIDERIUS horizontally, to additional scientific problems and data modalities beyond time series, and in depth, to more complex scientific tasks beyond ML development, including but not limited to data analysis, quantum control, experimental design, and detector monitoring.

The postdoctoral scholar will have opportunities to participate in projects associated with the DOE Genesis Mission and to collaborate with researchers at U.S. national laboratories, including Lawrence Berkeley National Laboratory, as well as other industry collaborators.

This role may be offered remotely, depending on the candidate’s work authorization restrictions and the PI's preferences.

Qualifications

Applicants must have, or expect to receive before the start date, a Ph.D. in Computer Science, Data Science, Physics, or a related field.

Experience with large language models and agentic AI is strongly preferred but not required. Strong programming proficiency and familiarity with using high-performance computing (HPC) systems are expected.

Experience in one or more of the following areas is desirable: machine learning, agentic AI, software engineering, scientific machine learning, or particle/nuclear physics. Candidates from both computational and physics backgrounds are encouraged to apply.

The position is a two-year initial appointment, renewable by mutual agreement. The expected start date is between January 7, 2027 and April 1, 2027. Candidates who are able to start earlier within this period will be preferred.

The expected salary range is $76,881–$82,677 per year, depending on experience and qualifications and in accordance with the UC Postdoctoral Scholar salary scale.

Application Materials

Applicants should submit the following through AcademicJobsOnline:

• Curriculum Vitae
• Research Statement
• Agent Statement — one page maximum
• Names and contact information for three references

Agent Statement

As part of the application, candidates should read the SIDERIUS paper:

Beyond a Better Score: Long-Horizon Agentic ML Development and Evaluation Protocol for Physics Time Series

Please click on shortened URL: https://tinyurl.com/yz95v6ac.

And prepare a one-page Agent Statement discussing one or both of thnulle following:

• Their assessment of SIDERIUS and how they would improve, enhance, or extend its performance and capabilities.
• How they would apply SIDERIUS to another scientific task, either in physics or in another scientific domain with which they are familiar.

The purpose of this statement is to understand how the candidate thinks about the future of agentic AI for science and how they would contribute new research directions.

Application Deadline

For full consideration, applications should be submitted by December 15, 2026. Applications received after December 15 may be reviewed until the position is filled, but candidates should apply by the first-round deadline to receive full consideration.

Contact

Questions regarding the position may be directed to:

Prof. Aobo Li
Halıcıoğlu Data Science Institute and Department of Physics

University of California San Diego

aol002@ucsd.edu