Climate Analytics Lab

Improving climate projections with machine learning

Location: University of California San Diego (La Jolla, CA)
Principal Investigators: Dr. Duncan Watson-Parris and Dr. Rose Yu

About the role

We are seeking highly motivated Postdoctoral Researchers to join an interdisciplinary initiative at the intersection of machine learning and Earth system science. The successful candidates will play a central role in developing the next generation of Earth system models (ESMs) by harnessing emerging agentic AI approaches to automate the discovery and improvement of atmospheric parameterizations.

A key component of this work is the recently Google-funded MOSAIC project, which combines observations, symbolic equation discovery and differentiable simulation, and rigorously evaluates and tests models against observations at scale using ClimateBench2. Beyond the core project, the role offers the flexibility to broadly shape the future of AI for climate science.

Position highlights

  • Guaranteed stability: a secure, 2-year guaranteed appointment.
  • Excellent compensation: a competitive salary commensurate with expertise.
  • Interdisciplinary collaboration: you will work directly with leading machine learning experts and climate scientists, bridging research across international modeling groups.
  • Work environment: a beautiful location, collaborating across the Scripps Institution of Oceanography and the Department of Computer Science and Engineering.

Key responsibilities

  • Develop agentic AI pipelines that autonomously discover and refine parameterizations within a differentiable ESM (JAX-GCM), learning from observational data across scales.
  • Leverage and curate cloud-hosted observational datasets (including CMIP and Obs4MIPs) to test model performance on historical changes and paleo-climate periods.
  • Define and test physicality protocols to ensure AI models conserve energy and obey physical relationships under unseen conditions.
  • Ensure research impact by adhering to open-science principles, making all assembled datasets, code pipelines, and publications publicly available.

Qualifications

  • Ph.D. in Computer Science, Climate Science, Physics, Data Science, or a closely related field. Experience with differentiable programming (JAX, PyTorch) is a plus.
  • Strong foundation in machine learning (e.g., deep learning, generative models, agentic/LLM-based systems) and/or Earth system modeling.
  • A demonstrated passion for addressing climate change and developing open, trustworthy AI systems.

How to apply

Please send a CV and a short research statement to dwatsonparris@ucsd.edu.

Use the subject line MOSAIC Postdoc Application (exactly as written) so your application is picked up. Clicking the email link above fills it in for you. Applications sent with a different subject line may be missed.

This is the only route for these positions. Please don’t email about them otherwise, and note that I am unable to reply to individual enquiries before applications are reviewed. Other routes into the lab are described on the Join the Lab page.