Project for a worldwide brand. Benchmarking of competitors' web platforms: reverse-engineering of infrastructure, development paradigms and technologies, analysis of performances and optimization (critical rendering path, runtime, memory footprint); short and long term technical tendencies and opportunities; recommendation on technologies, tooling and methodologies, design and three-years deployment roadmap of the next-gen platform; development of a proof of concept and masters.
Project for a famous luxury brand. Client iOS native application for tablet and smartphone developed with Objective-C and Swift; Node.js based socket and REST local servers for user registration, stock and orders management; PHP/MySQL based central server for data aggregation and analysis; content production, QA and distribution tools.
Client-based algorithm for processing webcam input with three subcomponents: ORB object detection for banknote value detection with JSFeat and predefined features descriptors; custom fast MSER algorithm for serial number digits blobs detection; OCR for serial number digits recognition based on Ocrad.js. Website awarded by a FWA.
Server-based algorithm deployed in two phases: a first one with a k-means clustering algorithm and relevance feedback, a second one with machine learning algorithm feeded with first phase collected data; desktop/mobile/SE optimized front-end with PageSpeed/YSlow 100% score.
Configuration-free task runner based on flow-based programming paradigm and local/FaaS modules that aims to provide some sort of Asset Pipeline for developers of static and Node.js based projects; database-free CMS layer for editors based on JSON-LD datastore and GIT versioning
Many other projects from native/web applications to offline tools, with social media APIs integration and asynchronous or real-time data flows for famous entertainemt and luxury brands.