Sobre

Project Description I own several retail stores across Brazil and need to automate the verification of cash stored in each store’s drawer. Currently, I perform video calls with each store to manually compare the system balance against the actual physical cash in the drawer. I need to automate this process: A mobile app (Android — priority; iOS optional) that captures images or short videos of the cash drawer, identifies and counts the banknotes (by denomination), and calculates the total cash. A backend service that receives this counting result, validates it, and reconciles with the store’s cash balance from my ERP via API. Automatic alerts when discrepancies exceed a set threshold. A simple web dashboard to visualize exceptions, history, and attach images/videos as evidence. Secure ERP integration via authentication (API key / OAuth2), with audit logs. Initial functionality: manual capture via photo and video-assisted capture (frame sampling). Future scope: automation with scheduled captures and daily report exports. I am looking for a professional (or team) with proven experience in computer vision/OCR for banknote detection, mobile development (Android), backend APIs, and secure system integrations. Knowledge of data security and automated testing is mandatory. Key Responsibilities Design and develop an Android app (minimum) to capture and process banknote images. Implement an algorithm/model (OpenCV + neural networks / Tesseract / custom ML) to detect and count banknotes by denomination. Build a backend (REST API) to receive counting results, reconcile with ERP balances, and store logs. Create a lightweight web dashboard for exception management and reporting. Document endpoints, workflows, and provide end-user documentation. Deliver unit, integration, and field tests (sample images/videos will be provided). Version control on Git (with PR workflow) and deployment documentation (Docker preferred). Mandatory Requirements Proven experience with computer vision / OCR projects in real-world environments with variable lighting. Experience in Android native development (Kotlin/Java). Flutter/React Native acceptable if justified. Backend experience: Node.js/Express, Python/Flask/FastAPI, or Java/Spring. API integration expertise (OAuth2 or API key), working with JSON. Security best practices: TLS, secure credential storage, audit logs. Portfolio or previous similar projects (screenshots, demos, or code). Good communication in Portuguese or English. Nice-to-Have iOS development (Swift) or cross-platform (Flutter/React Native). Experience with ML models (TensorFlow/PyTorch) for object detection. On-device inference (TensorFlow Lite) to reduce latency and preserve privacy. Infrastructure skills (Docker, Kubernetes, CI/CD).