System
Capabilities

Interfaces that hold up under real traffic. I build component-driven frontends with type safety by default, and I care about the last mile — rendering performance, accessibility, and how it feels under the thumb.

The logic core. Spring Boot is home turf — MVC architectures, REST APIs, background pipelines — but the toolbox spans JVM, Python and native mobile. Python carries the ML and computer-vision work: inference services, data and image pipelines, automation. Swift carries the native side, including Core ML models running on-device rather than round-tripping to a server. Data lives in whatever store actually fits the access pattern.

The operator mindset, in practice: a decade of virtualization, seven-plus years running Linux and Windows fleets, and a personal datacenter that's treated like production — containerized, monitored, reverse-proxied, and reachable through zero-trust remote access.

Models put to work, not worshipped. Production chatbots on hosted frontier models, plus a full self-hosted inference stack on the homelab's GPU node — model serving behind an OpenAI-compatible gateway, a separate embedding service, and a vector store for retrieval-augmented generation. Nothing leaves the rack unless I want it to. Alongside it, computer vision: CUDA-accelerated pipelines where classical OpenCV preprocessing feeds the neural stage, and on-device inference shipped to Apple hardware through Core ML. Then the analytics side: data models that turn raw SQL into decisions.

The creative layer that predates the code: eight-plus years of video editing, seven of audio, and a working photographer's eye. This is what runs the production booth at GDG Cebu events — and what keeps the interfaces honest.