Accessible mobile engineering

Volunteer software engineering for Petanetra — an accessibility-oriented product that combines on-device object detection with offline, multilingual spatial guidance. My work crossed Android architecture, applied-ML integration, product instrumentation, and cross-platform delivery.

Focus

• On-device ML integration — integrated object detection with ONNX and TensorFlow Lite inside a maintainable Android architecture rather than a one-off demo.
• Architecture — applied clean-architecture principles so the ML capability had clearer boundaries and could evolve without turning the application into a tightly coupled prototype.
• Product behaviour and instrumentation — worked on offline, multilingual spatial guidance and refactored attribution tracking across onboarding and feature navigation, adding properties for more robust, accurate, and granular product signals.
• Team coordination — served as Android team lead and, from January 2025, took broader project-coordination responsibility by breaking longer-term cross-platform work into smaller pieces that volunteer contributors could own.

The team worked within offline constraints and used product feedback from real users while the mobile and applied-ML pieces were developed. Volunteer delivery also made explicit boundaries, dependencies, and manageable work slices important to keeping progress moving.