Exploring philosophy and computation
See Details Below
I'm a first-year student at UC Berkeley studying Computer Science with particular interests in reasoning & inference in AI systems.
I seek to understand and create complex systems through computational and theoretical frameworks.
My work spans computer vision, natural language processing, and computational theories that reveal pattern and chaos.
See my CV here.
Oct 2025 - Present
Collaborating within Keenan’s Group to architect a multi-agent AI framework designed for individual tree labeling and uncertainty classification & quantification in biomass estimation.
Engineering multimodal pipelines to process and interpret spatial data, including 3D point clouds and raster data.
Applying multi-step reasoning workflows and prompt design to automate the classification of structural complexity in forest canopy datasets.
Oct 2025 - Present
Investigated gender bias in LLMs using NLP analysis on gender-blind résumés, employing prompt-sensitivity testing and vertical comparison between DeepSeek-R1 and DeepSeek-V3 models.
Designed a methodological framework to isolate the effect of alignment post-training, revealing that gendered framing persists as latent encoding in representation space despite surface-level behavioral constraints.
Demonstrated that bias is reinforced through linguistic form, producing novel insights into the limitations of alignment training and contributing to broader discourse on stereotype consolidation in language models.
Read the paper here.
Sep 2025 - Present
Developing non-radiographic, home-based scoliosis detection, tracking & rehabilitation platform.
Engineering a Computer Vision Pipeline using OpenCV to process user-submitted scans, reconstruct high-fidelity digital twins, and extract critical spinal geometry for model prediction.
Deploying full-stack, cross-platform applications.
May 2022 - Sep 2024
Developed statistical learning pipelines using a Linear Mixed Model in MATLAB to perform high-dimensional, covariate-controlled analysis on a large-scale neuroimaging dataset. Applied graph theory algorithms to model structural covariance. See the publication here.
May 2024
Engineered an end-to-end NLP pipeline using Python, NLTK, and scikit-learn to preprocess 450k+ text data (tokenization, lemmatization, stop-word removal) and extract features via TF-IDF vectorization.
Proactively enhanced model robustness by curating training sets to address complex NLP challenges (sarcasm, slang, and multilingual text).
Apr 2023 - Jul 2024
Led China's first Grand Slam delegation, analyzing particulate hazards and designing defense systems based on particle-matter interaction models.
On Gendered Framing as a Latent Encoding in Technical Self-Representation and Evaluation
View on GitHub →fMRI Data Analysis using Graph Theoretical Algorithms and ML
See Publication →