ÐÓ°ÉÐÔÊÀ½ç

My research focuses on making artificial intelligence, especially large language models (LLMs), fairer and more socially responsible. These models are now used in education, hiring, healthcare, and everyday communication, but they can unknowingly reproduce social, cultural, or gender biases from the data they learn.

My work identifies these hidden biases and designs ways to measure and reduce them, ensuring AI systems treat people from all backgrounds equitably. By studying how LLMs respond to factors such as nationality, age, and culture, my goal is to build AI tools that understand human diversity and support more just, trustworthy technologies for society. 

Tasmanian devils are the largest living marsupial carnivores and are found exclusively on the island of Tasmania. Devils have been afflicted with a transmissible cancer, devil facial tumor disease (DFTD), which was discovered nearly thirty years ago. DFTD spreads when an uninfected devil bites the tumor mass of an infected devil and is uniformly fatal; DFTD has thus led to island-wide devil population declines of 80%. Although early epidemiological models predicted extinction, devil populations have persisted and sometimes rebounded in the face of DFTD. Recent evidence has revealed rapid evolution of both devils and DFTD, indicating the potential for continued devil existence. I use genomic tools to study devil and DFTD evolution to help predict the trajectory of this epidemic, thus aiding conservation efforts and gaining insight into cancer and pathogen evolution.

Despite evidence of devil and DFTD evolution, previous studies had not yet looked for the possibility of a reciprocal evolutionary process called coevolution. To address this gap, we used a recently developed statistical model to look for genomic signatures of coevolution, investigating the contribution of devil-DFTD genomic interactions in explaining how quickly devils become infected with DFTD. We found that devil-DFTD genomic interactions explained roughly three times more variation in infection rate than either devil or DFTD genomes alone. We also found that the interacting genes between devils and DFTD were enriched for cancer function, and interactions between specific mutations in devil and DFTD genes were associated with differing infection rates. Our results strongly suggest that coevolution is important to the evolutionary trajectory of devils and DFTD, emphasizing the importance of such interactions when considering devil management strategies. Furthermore, the interacting genes we found represent potential candidates in future research of human cancers, and the framework we used to investigate genomic interactions is broadly applicable to human pathogens.

My research focus in the Staruschenko Lab encompasses 3 topics. The first project investigates how opioids affect kidney function. Despite the severity of the present-day opioid crisis, little is known about how opioids influence kidney health. Through a combination of animal model and cell studies, we have elucidated that upon exposure to kappa opioids activates the kappa opioid receptor on podocytes, leading to the opening of the TRPC6 channel and causing calcium influx overload. This calcium overload results in podocyte damage and leads to pathologies such as albuminuria and hypertension. This mechanism provides a possible target to where, if blocked, could protect the kidneys of individuals using opioids. 

My second investigation examines how cannabidiol (CBD) affects podocytes in vitro. Here, we established that prolonged exposures to high levels of CBD causes mitochondrial dysfunction, altered calcium handling and increased cell viability. Because CBD is widely used yet highly underregulated and understudied, our aim is to provide greater insight into its action in the kidney. 

Lastly, my third project explores how semaglutide affects kidney function in the adolescent population. With the exponential rise of adolescent obesity and the increasing production of weight loss medications, we aim to determine the side effect profile and precise mechanisms of the glucagon like peptide-1 (GLP-1) agonist, semaglutide, in the kidney. Thus far, we have found that semaglutide use in adolescent male mice does not significantly affect parameters such as lean mass and bone density, and it marginally improves glomerular filtration rate. Furthermore, we demonstrated that semaglutide treatment upregulated pro-mitochondrial function and autophagy proteins, while downregulating pro-inflammatory proteins in the kidney cortex. This work addresses a critical knowledge gap in the weight loss field and will serve as my thesis project.