XETOR: AI-Powered Digital Waste Bank & Green FinTech Mobile Ecosystem
Native Android green fintech application integrating AI vision scanning and e-wallet mechanics to convert sorted waste into real financial value.
Overview
XETOR tackles low household waste separation rates caused by complex manual drop-offs and absence of tangible financial incentives. Aligning directly with UN SDG 12 (Responsible Consumption and Production), this platform bridges AI computer vision image classification with an integrated financial wallet. Users scan materials for instant categorization, hand over recyclables to affiliated collection partners via QR validation, and seamlessly redeem earned digital reward points into cash withdrawals, transit passes, or utility bill settlements.
Challenges & Solutions
Merging client-side computer vision inference with low-latency financial transaction processing presented notable engineering challenges:
Optimizing camera image classification inference latency in Android to achieve real-time classification without excessive memory footprint.
Ensuring transactional consistency across atomic point-to-cash conversion queries during concurrent high-volume partner drop-off events.
Preventing physical deposit fraud through dynamic time-bound QR code tokens featuring cryptographic expiration timers.
Key Responsibilities
Architected and built the complete native Android mobile application leveraging modern Kotlin and Jetpack Compose declarative UI
Developed high-throughput backend API microservices utilizing Go (Golang) backed by a normalized PostgreSQL database
Integrated on-device AI Vision classification pipelines enabling instantaneous recyclables recognition and sorting guidance
Engineered full-cycle fintech transaction modules covering point-to-cash conversions, peer transfers, and bank withdrawals
Built a secure drop-off verification system leveraging dynamic timed QR codes with real-time countdown validation for collection partners
Result & Outcomes
The functional production prototype and backend architecture of XETOR demonstrated measurable viability for green fintech systems:
Delivered a production-grade native Android client encompassing 10+ reactive screens utilizing clean Jetpack Compose architecture.
Achieved sub-100ms API response latencies on Go microservices handling high-frequency e-wallet operations and QR authorizations.
Validated a sustainable circular economy model directly addressing UN Sustainable Development Goal 12 benchmarks.
Lessons Learned
Architecting XETOR delivered critical insights into pairing mobile camera sensor pipelines with secure financial ledgers:
Centralized reactive state management in Jetpack Compose is vital to synchronize live wallet balances and camera states without redundant recompositions.
Financial reward applications require server-enforced double-entry verification rather than relying on client-side state assertions.
Technologies Used
Kotlin
Jetpack Compose
Go
Key Features
AI Waste Scanner — Computer vision camera feature accurately identifying recyclable material categories with contextual disposal guidance.
QR Drop-Off Verification — Rapid physical handover validation utilizing dynamic, time-sensitive QR codes scanned by verified collection partners.
Integrated Green E-Wallet — Full-featured financial wallet handling dynamic point-to-cash conversions, peer transfers, and instant bank withdrawals.
Multi-Utility Redemption — Reward redemption network enabling users to spend balances on eco-friendly marketplace items, public transit, and utilities.
Eco-Impact Analytics — Sustainability dashboard tracking measurable user environmental impact, calculating energy saved (kWh) and CO2 avoided (kg).
Partner Drop-Off History — Comprehensive ledger tracking historical drop-off timestamps, verified weight measurements, and generated reward yields.