Analyzed trade-off between speed, GPU parallelism, and energy cost. Quantified that scaling GPUs for faster inference increases energy use by up to 55%, giving KPMG a clearer framework for evaluating when performance gains justify added compute.

Built an AI caregiving copilot that transforms fragmented family updates, care routines, and patient needs into actionable coordination support. Focused on context-aware assistance, reduced caregiver burden, and AI that fits into real family care workflows.

My first full-stack project (2023): Baybayin is a pre-colonial Philippine writing system that is no longer widely used, and this simple but meaningful project was my first attempt at turning cultural inspiration into a working end-to-end application. Tested by 35 users.

Implemented a limit order book in C++ to model matching logic, order flow, and market data structures.

Do you know where your 24 hours go? Designed simple logging and visual breakdowns to help users spot patterns in how they spend their day.

Featured by CWRU for International Women and Girls in Science Day 2026.

Featured for work with the West Side Market on market strategy, customer insights, and renovation-related recommendations.

Compared ICA and EMD for denoising simulated EEG time series, studying how stochastic and deterministic algorithms remove artifacts from electroencephalogram signals. This was my IB Computer Science extended essay: an early independent research project in signal processing and computational experimentation.

Analyzed prism-adaptation experiments with 40+ participants to test whether motor learning transfers between different throwing movements. Led the Python pipeline for cleaning displacement data, visualizing adaptation patterns, and fitting exponential decay curves to estimate error-correction behavior.