Developing Machine Learning Tools for Low-temperature Stellar Spectral Model Fitting
Project Description:
High quality spectroscopic data for the lowest-mass stars and brown dwarfs are being collected by JWST, Euclid, and SPHEREx space telescopes, opening up new opportunities to study the atmospheres of large populations of low-temperature objects. In this project, you will be explore, develop, and test new machine learning tools to analyze these spectra, with the goal of improving the speed and precision in measuring the physical properties of our nearest stellar and substellar neighbors.
Skills required/recommended:
prior course in astronomy preferred, some programming experience (python will be part of the program)
Helpful Resources:
Zhou et al. (2025): "Classifying Cool Dwarfs: Comprehensive Spectral Typing of Field and Peculiar Dwarfs Using Machine Learning" https://ui.adsabs.harvard.edu/abs/2025ApJ...992...93Z/abstract
Lueber & Burgasser (2025): "Comparing Grid Model Fitting Methodologies for Low-temperature Atmospheres: Markov Chain Monte Carlo versus Random Forest Retrieval": https://ui.adsabs.harvard.edu/abs/2025ApJ...988...31L/abstract
Draxl Giannoni et al. (2026): "Identifying and Characterizing Very Low Mass Spectral Blend Binaries with Machine Learning Methods" https://ui.adsabs.harvard.edu/abs/2025arXiv251212098D/abstract
A rare brown dwarf in an extreme radiation environment
Project Description:
Brown dwarfs are useful analogs to giant planets because they share similar atmospheric temperatures, chemical compositions, and physics. LP 261-75 C is a brown dwarf on a rare, short-period (1.88 day) orbit around an extremely active low-mass star, where stellar activity likely drives photochemistry and atmospheric structure of the brown dwarf. In this project, a student will help analyze spectroscopic observations of LP 261-75 C from Gemini-North/IGRINS-2. These observations aim to characterize the chemical and temperature structure of the day and night-side atmosphere of LP 261-75 C to explore the influence of radiation from extreme stellar activity. Student activities may include atmospheric modeling of the brown dwarf, and/or characterization of energetic flares from the host star LP 261-75 A to support analysis of spectroscopic observations.
Skills required/recommended:
Some background with python programming is recommended
Helpful resources:
Gibbs & Fitzgerald (2025):
"Simulations of Flare Chemistry in Brown Dwarf Companions to Active M Dwarfs": https://doi.org/10.3847/1538-4357/ae0477
Irwin et al. (2018): "Four New Eclipsing Mid M-dwarf Systems from the New Luyten Two Tenths Catalog": https://doi.org/10.3847/1538-3881/aad9a3
Characterizing post-common envelope white dwarf binaries in star clusters
Project Description:
Many stars live in binary systems and can go through a short but dramatic phase called common envelope evolution, during which the two stars briefly share a single envelope and their orbit rapidly shrinks. This process is thought to play a key role in forming compact object (white dwarf, neutron star or black hole) binaries, but it is poorly understood because there are so few observed and characterized examples of post-common envelope binaries. In this project, we will analyze a newly discovered sample of post-common envelope white dwarf binaries in star clusters and characterize their properties using observational data. Studying these systems in star clusters, where the stellar ages are known, helps us better understand how common envelope evolution shapes the binaries we observe today and many Type Ia supernovae and gravitational waves we will observe in the future.
Skills required/recommended:
Basic experience in Python programming is preferred.
Helpful resources:
Grondin et al. (2024): "Discovering post-common envelope binaries in clusters" : https://ui.adsabs.harvard.edu/abs/2024ApJ...976..102G/abstract
Investigating the Atmospheric Compositions of Brown Dwarfs and Low Mass Stars
Project Description:
Spectroscopy is a vital tool for characterizing the atmospheres of extra solar planets and stars, which can in turn yield information about the origins, chemistry and evolution of these systems. Brown dwarfs, with masses between those of exoplanets and stars, possess temperatures and gravities that overlap with the population of extrasolar gas giant planets. Studying brown dwarf atmospheres therefore offers a promising path forward in understanding the properties of giant planet atmospheres. Similarly, characterizing the atmospheres of low mass M dwarf stars, the most common planet host star in the galaxy, is critical to understand the planets they host. In this project, the mentee will analyze spectroscopic observations of brown dwarfs or M dwarf atmospheres, depending on interest, from the FIRE spectrograph on the 6.5 m Magellan telescope. They will then use modeling techniques to constrain their atmospheric compositions.
Skills required/recommended:
Python (part of summer training)
Helpful resources:
Madhusudhan (2019): "Exoplanetary Atmospheres: Key Insights, Challenges, and Prospects": https://ui.adsabs.harvard.edu/abs/2019ARA%26A..57..617M/abstract
Uncovering the origin of gravitational wave sources
Project Description:
Over the past few years, the groundbreaking detections of gravitational wave signals from merging binary black holes and neutron stars by LIGO/Virgo have opened a new window to the cosmos. One key question regarding these gravitational wave sources concerns the nature of their origin. Dynamical formation in dense stellar environments like globular clusters has emerged as an important formation channel, corroborated by recent numerical simulations and observational indications showing globular clusters contain dynamically significant populations of stellar-mass black holes throughout their lifetimes. For this project, we will use N-body simulations of globular clusters to investigate the formation of black hole binary mergers in these systems. The student will also work closely with Professor Floor Broekgaarden at UCSD.
Skills required/recommended:
Basic experience in computer programming (python preferred).
Helpful resources:
I will pass along more materials before the summer begins!
Forming binary black holes in EAGLE
Project Description:
Mergers of binary black holes depend strongly on the metallicity of the massive stars from which they formed. Therefore, the assumptions for the star formation history used in modeling populations of binary black holes can have a significant impact on the resulting population. This project will explore the star formation and metallicity history of the EAGLE cosmological simulations and use it to model a binary black hole merger population. This project will utilize both cosmological simulations (EAGLE) and binary population synthesis simulations (COMPAS).
Skills required/recommended:
Some experience in Python, interest in stellar evolution and gravitational waves
Searching for Circumbinary Planets in TESS Data
Project Description:
NASA's TESS mission (Transiting Exoplanet Survey Satellite) has been providing photometric data for bright stars across most of the sky for the past seven years. The primary mission of TESS is to look for transiting exoplanets, but owing to the nature of the data many other types of science can be done. For the STARTastro program, the student will work with the mentor and search for transiting circumbinary planets (these are planets that orbit both stars in a binary system). Over the years a large catalog of eclipsing binaries (these are binaries whose orbits are seen nearly edge-on from Earth and as a consequence the two component stars periodically eclipse each other) has been constructed, and the light curves of these eclipsing binaries need to be inspected for the characteristic transits that could indicate a circumbinary planet. In addition, if time permits, other projects involving eclipsing binaries can be done, such as finding the masses and radii of the component stars in newly discovered systems.
Skills required/recommended:
Coding is a helpful skill.
Helpful resources:
Learn about MAST and the STScI Archive: https://archive.stsci.edu/
Understanding the Interstellar Medium in Nearby Dwarf Galaxies
Project Description:
Dwarf galaxies are fundamental building blocks of larger galaxies, yet their interstellar medium (ISM) remains complex and dynamic. These galaxies are characterized by high HI content but low dust and metal abundances, making them key laboratories for studying star formation and chemical evolution in metal-poor environments. In this project, the student will analyze archival JWST, Keck, and VLA observations to investigate the dust, metal, and gas content of nearby dwarf galaxies. This work will help answer broader questions, such as how dwarf galaxies evolve chemically and what factors regulate their ISM properties. Potential avenues for exploration include comparing dust-to-gas ratios, examining metallicity gradients, and investigating neutral & ionized gas kinematics.
Skills required/recommended:
I recommend having familiarity with Python programming. The student will use Astropy packages and libraries to carry out several aspects of this project. In addition, they will use codes like CIGALE, Source Extractor.
Helpful resources:
Messa et al. (2018) "The young star cluster population of M51 with LEGUS - I. A comprehensive study of cluster formation and evolution": https://ui.adsabs.harvard.edu/abs/2018MNRAS.473..996M/abstract
Shabani et al. (2018) "Search for star cluster age gradients across spiral arms of three LEGUS disc galaxies": https://ui.adsabs.harvard.edu/abs/2018MNRAS.478.3590S/abstract
Turner et al. (2021) "PHANGS-HST: star cluster spectral energy distribution fitting with CIGALE": https://ui.adsabs.harvard.edu/abs/2021MNRAS.502.1366T/abstract
Thilker et al. (2007): "A Search for Extended Ultraviolet Disk (XUV-Disk) Galaxies in the Local Universe": https://ui.adsabs.harvard.edu/abs/2007ApJS..173..538T/abstract
X-Ray Instrumentation Development for STROBE-X
Project Description:
This research opportunity focuses on the testing, validation, and characterization of mixed-signal electronics readout systems for silicon drift detectors (SDDs), directly supporting the development of instrumentation for an all-sky monitoring X-ray space telescope.
Participants will acquire hands-on experience in precision electronics soldering and rework, circuit testing and troubleshooting, application-specific integrated circuit (ASIC) characterization, and scientific data analysis using Python. In addition, the project provides foundational training in X-ray astrophysics and front-end readout design for high-energy astrophysics instrumentation.
Skills required/recommended:
None
Helpful resources:
STROBE-X webpage: https://strobe-x.org/Instruments.html
Unsupervised Machine Learning Applied to JWST Spectral Maps of Nebulae
Project Description:
The James Webb Space Telescope has two integral field spectrographs, NIRSpec and MIRI-MRS. These provide spectral cubes with spatial and wavelength dimensions. Due to the incredible sensitivity of JWST, even NIRSpec and MIRI-MRS observations of nebulae in nearby galaxies detect a huge number of spectral features, including emission lines, vibrational emission features from small dust grains, and absorption from dust and ices. Dealing with this huge amount of data is daunting. Over the summer, we will try a new approach to analyzing these spectral cubes, using an unsupervised machine learning technique called “non-negative matrix factorization”. By identifying a small set of components that represent most of the variation in the observations, we can separate different part of the interstellar gas and dust and see how they vary. We will identify the principal spectral components in the nebulae and compare them to known lists of emission lines. The project will start with observations of the N13 nebula in the Small Magellanic Cloud and, depending on what we learn, expand to other regions and galaxies.
Skills required/recommended:
no prior skills necessary, but we will use python and jupyter notebooks and astropy a lot
Helpful resources:
Berné et al. (2022): "PDRs4All: A JWST Early Release Science Program on Radiative Feedback from Massive Stars": https://iopscience.iop.org/article/10.1088/1538-3873/ac604c
The Chemistry of Colors on Jupiter's Ocean Moon Europa
Project Description:
Europa’s geologically young, disrupted surface is painted by various hues of red, yellow, and brown. Such colors relate to salts and sulfurous material that can help us understand the composition of Europa’s interior and the alteration of its surface chemistry by sulfur and other charged particles ultimately derived from the volcanic emissions of nearby Io. In fact, visible-wavelength spectroscopy from the Hubble Space Telescope (HST) has revealed a number of geographically varying spectral features, including a so-called “color center” absorption indicating the presence of irradiated NaCl from the internal ocean. Studying how the colors on Europa correlate with different geologic terrains and patterns of charged particle bombardment will be key to understanding the chemistry of both endogenic (sourced from the interior of Europa) and exogenic (sourced from outside Europa) materials on Europa’s surface. This project will revisit spatially resolved ultraviolet and visible-wavelength spectra of Europa from HST, with the goal of further constraining the chemistry underlying the colors we see in imagery and revealing how it relates to different geologic terrains and exogenic effects. The project will involve data reduction, mapping, endmember identification, and comparison with laboratory data.
Skills required/recommended:
Basic introduction to scientific coding (e.g., Python, matlab, etc.) would be beneficial. Basic introduction to spectroscopy and/or chemistry helpful, but not required.
Helpful resources:
I will provide resources closer to the start date or at the start of the first week.
Find the Planet!
Project Description:
You will investigate data taken with the 40-inch Nickel Telescope at Lick Observatory to try and find transiting planets across multiple observing nights. All observations include a planetary transit, but the main goal will be understanding the systematics in the data to actually detect the planet transit, and identify the mid-point of the transit. You will learn the basics of CCDs, data reduction, and high-precision photometry. There is the potential for hands-on observing experience using the Nickel telescope as part of this project.
Skills required/recommended:
Helpful resources: