EESA Research Projects Summer 2026

Soils store a huge amount of Earth’s carbon, but we still do not fully understand what makes some carbon stick around while other carbon disappears. This project will test how fresh microbial remains compare to recycled microbial remains, and how strongly each one attaches to different soil minerals. The student/visiting faculty member will help run lab incubations and mineral binding tests, track carbon changes over time, and organize results into clear figures. Outcomes include a better picture of what controls long-term soil carbon storage, plus hands-on skills in experimental design, basic lab methods, data visualization, and communicating results.  

Desired Education Level: Community college student, Undergraduate student, Visiting Faculty member

Relevant Internship Programs: CCI, SULI, VFP

Mentor: Romy Chakraborty (rchakraborty@lbl.gov) 

Viral diseases threaten California vineyards by reducing yield and fruit quality, often long before symptoms are visible. This proposed project will utilize a handheld field spectrometer, a machine-learning model to predict early signs of viral infection, and a simulation of canopy reflectance to translate results into clear, actionable insights for growers. The VFP researcher, with a student, will collect field data, analyze patterns, and develop prediction models to aid in the timely detection of viral infections and mitigate the spread. Outcomes include improved vine virus prediction models for vineyard management, as well as hands-on skills in data collection, plant health monitoring, and the application of AI/LLM in agriculture.

Desired Education Level: Visiting Faculty member

Relevant Internship Programs: VFP

Mentors: Nicola Falco (nicolafalco@lbl.gov) 

TBD

Desired Education Level: Undergraduate student

Relevant Internship Programs: SULI

Mentors: Gianna Marschmann (GLMarschmann@lbl.gov)  

This project addresses critical location inaccuracies and data gaps in existing oil, gas, and geothermal well databases that hinder effective environmental remediation and the safe use of fossil fuels. The research combines precise field data collection with advanced Artificial Intelligence (AI) to validate well infrastructure integrity. The core of the project involves deploying AI-powered Computer Vision and Optical Character Recognition (OCR) models, along with a comprehensive suite of more than a hundred Large Language Models (LLMs), to extract, analyze, and consolidate vital operational data from weathered, rusted, and degraded physical signage at well sites.

The overall goal of the project is to develop a scalable, automated workflow that can update Federal or State databases with verified field data and identify orphan wells that require plugging or remediation. This ground-truth data is essential for calibrating aerial and remote sensing surveys (e.g., magnetometry, LiDAR, radar, PolSAR, and multi- and hyperspectral sensors) as well as for developing data fusion techniques and training future machine learning models. This supports the Office priority of Advanced Oil and Gas Production Technologies by modernizing how legacy assets are tracked. Key learning outcomes include a comprehensive evaluation of AI performance on industrial assets, the development of robust data-extraction techniques for challenging field environments, and the quantification of the data improvements required for accurate characterization.

Desired Education Level: Community college student, Undergraduate student, Masters student, PhD student, Visiting Faculty member

Relevant Internship Programs: CCI, MLEF, VFP, DOE Science Graduate Student Research (SCGSR)

Mentor: Andre Santos (ALDSantos@lbl.gov) 

Belowground microbiome, including both bacteria and fungi, play key roles in soil biogeochemical processes, and beneficial microbiome has been reported to enhance plant growth in various environmental conditions, which is a promising solution for bioenergy and agricultural production and soil health. In Chakraborty Lab, we are developing genome-resolved synthetic microbial communities and mechanistically investigating their effects on the growth and physiology of different plant species, tracking their fate among native microbiome, and developing tools to non-destructively evaluate the plant health across time and treatment groups. The participant(s) will learn a broad range of lab assays including bacterial culture, DNA extraction, assays for plant and bacterial phenotypical and physiological traits, and biological statistics. 

Desired Education Level: Community college student, Undergraduate student, Visiting Faculty member

Relevant Internship Programs: CCI, SULI, VFP

Mentors: Yuguo Yang (yyang14@lbl.gov)  

The project seeks to develop an Agentic system for the International Land Model Benchmarking (ILAMB) system. The agent system will advance the science discovery of RUBISCO project, and provide insights to land biogeochemistry and land-atmosphere interactions.

Desired Education Level: PhD student

Relevant Internship Programs: 

Mentors: Qing Zhu (qzhu@lbl.gov)   

Software is increasingly becoming interoperable such that we can mix and match components to rapidly develop new models. In this project, we aim at developing and/or demonstrating software interfaces that enable this interoperability for geochemical models. The products of this work will be used by researchers worldwide to develop the next generation of multiphysics simulators. This project provides an opportunity to hone your Python programming as well as your scientific communication skills.

Desired Education Level: Community college student

Relevant Internship Programs: CCI

Mentor: Sergi Molins (smolins@lbl.gov)  

Data-center operators require 99.999% energy availability, while next-generation geothermal power can provide roughly 90%. Can high-temperature heat storage in depleted oil fields compensate for the shortfall with sufficient frequency to meet data-center needs?

To address this question, you will (Task 1) analyze the frequency of peak-power demand in data centers and (Task 2) numerically simulate the discharge behavior of a simplified high-temperature heat-storage system in a depleted oil field to assess the degree of matching.

The work will be carried out using the TOUGH simulation codes (https://tough.lbl.gov).

Desired Education Level: Undergraduate student, Masters student, PhD student, Visiting Faculty member

Relevant Internship Programs: VFP

Mentor: Eva Schill  (eschill@lbl.gov)  

This project will use autonomous sample preparation and characterization to discover minerals for subsurface energy storage. The successful candidate will learn how to plan and execute experiments programmatically; how to use AI to acquire and process data; and how to do hypothesis-driven AI-guided research. 

Desired Education Level: Community college student, Undergraduate student, Masters student, PhD student, Visiting Faculty member 

Relevant Internship Programs: SULI, BLUR, CCI, MLEF, GEM, DOE Science Graduate Student Research (SCGSR), VFP 

Mentors: Mike Whittaker (mwhittaker@lbl.gov)

This summer project tackles a key challenge for wind turbines, pipelines, and other energy infrastructure: catching hidden concrete damage early, before cracks let in water and salt that can cause corrosion and costly failures. The student will build and test a portable sensing device that “maps” where concrete is wet, salt-contaminated, or degrading, and then add onboard machine-learning so the system can automatically interpret measurements in the field instead of relying on a full lab setup. By the end, the project aims to deliver a compact prototype that can flag high-risk areas and guide maintenance decisions, especially in remote or harsh environments. The student will gain hands-on skills in hardware prototyping, experimental testing, data analysis, and practical ML deployment—learning how to turn real sensor signals into actionable engineering insight. 

Desired Education Level: Visiting Faculty member 

Relevant Internship Programs: VFP 

Mentors: Yuxin Wu (ywu3@lbl.gov) 

The focus of the project is on using AI/ML algorithms, in particular, Graph Neural Networks. The project will allow participants to become familiar with AI/ML approaches for developing interaction potentials for gas-phase molecules with varying isotopic compositions. For example, by using the ML-potential developed for simple organic molecules, we will predict isotopic effects in larger (bio)molecules that cannot be studied directly with high-level quantum chemistry methods. Finally, the developed potentials will be used to model isotopic exchange and fractionation reactions, providing insight into atmospheric chemistry related to energy-production processes. 

Desired Education Level: Community College student, Undergraduate student, Masters or PhD student, Visiting Faculty member

Relevant Internship Programs: MLEF, DOE Science Graduate Student Research (SCGSR), VFP

Mentor: Piotr Zarzycki (ppzarzycki@lbl.gov)