Academic work and technical explorations at the intersection of machine vision, efficient computation, and secure system boundaries.
Notes from the frontier.
03 SELECTED WORKS
01
Undergraduate thesis 2023
Computer vision
Android
ResNet50
Vegetable Identification & Quality Classification
An Android application evaluating carrots, potatoes, okra, eggplant, and bitter gourd. The study achieved 100% vegetable identification accuracy and 95.38% export quality classification accuracy.
A secure, high-efficiency domain-specific architecture. Deterministic systolic array pipelines address dark silicon power limits and speculative side-channel attacks.
Hardware-level isolation in multi-tenant environments. Exploring secure microarchitectures and domain-specific sandboxes that prevent speculative leakage without compromising execution speed.
My current interests connect efficient edge computation with predictable execution and secure system boundaries, building on computer vision work and graduate studies at Mapúa University.
For a manuscript, implementation details, or a discussion of reported thesis results, contact me directly.