Research & Insights
AI that helps people do more
I have always been drawn to what AI can do for people when it gives timely, useful feedback. That curiosity started with my thesis — a mobile app that helps Filipinos learn Filipino Sign Language by checking whether hand shapes match the FSL alphabet — and now shapes how I build automations and skills for AI coding agents.
The thesis and the skills do not share an architecture. They share one idea: useful AI shows you whether you are on track, right when it helps. Seven guides · MIT · GitHub holds the current manuals.
From FSL feedback to timely agent feedback
One publication, one idea — useful AI gives people feedback while they can still act on it.
Handsy: Alphabet Sign Language Tutorial Mobile Application with Hand Recognition
Lead researcher & developer · Computer vision + neural networks for Filipino Sign Language
FSL alphabet tutorial that uses computer vision and neural networks to recognize finger alphabets — built to support inclusive learning for the PWD sector.
Learning FSL finger alphabets is harder without immediate, consistent feedback. Learners need a way to practice and know if their hand shapes are correct.
Handsy combines computer vision and neural networks to detect and classify hand shapes in real time, paired with guided tutorial flows that help learners practice the FSL alphabet with feedback.
Led research and development as lead researcher and developer; co-authored the published study.
- Sign — learner performs an FSL alphabet hand shape
- Recognize — computer vision and neural network classify the shape
- Respond — app shows whether the sign is correct and guides the next try
- Learn — repetition with immediate feedback builds consistent FSL practice
Open-source skills that help coding agents ship better work
Seven guides that help you plan work, deliver it in separate pull requests, finalize it after review, and keep project guidance readable. Each guide shows fit, workflow, setup, and limits.
/unslopify
/document-for-agents
/document-for-humans
/plan-this
/implement-this
/review-this
/fix-this
Current Focus Areas
Four areas I build and study — each grounded in real workflows and current hands-on exploration, distinct from published research. The skills above are the hands-on proof of how I work with agents daily.
Practical AI Agents
Assistants that read, route, and act across your tools — scoped to real tasks, not demos. Current capability and focus area.
Workflow Automation
Automations that connect websites, spreadsheets, CRMs, and APIs to reduce manual handoffs and missed follow-ups.
AI-Assisted Web Development
AI-assisted builds that keep delivery fast without sacrificing clarity, performance, or maintainability.
WordPress Engineering
Custom templates, Secure Custom Fields, Greenshift UI, and Cloudflare-hardened WordPress systems teams can manage.
Research, code, and selected work
Technical work and code on GitHub, professional updates on LinkedIn.