unicsfscker

Led the development and implementation of the frontend architecture for an AI editor, using Astro for high-performance landing pages and React (with TypeScript and Redux/Redux-Saga) for the core interactive application, ensuring flexibility and high speed; Solved performance issues when handling large GIF animations by implementing frame parsing, Web Workers, and smart caching; Ensured reliable local storage of user data (projects, settings) by implementing a cache in IndexedDB, guaranteeing instant access and offline capabilities; Created an intuitive and pixel-perfect masking tool using the Canvas API and complex geometric calculations, which became a key feature for the AI regeneration function; Integrated AI models, working closely with the AI team to define the API contract; Utilized Redux and Redux-Saga for managing asynchronous AI operations and numerous interactive user actions.

Optimized Next.js landing pages: increased SEO and Lighthouse performance scores by 5-6 points by implementing middleware redirects, lazy loading, and accessibility improvements; Removed jQuery and its related dependencies, refactoring code sections to pure React, which reduced the bundle size by 3-4% and accelerated page load times; Implemented lazy loading for admin panel components and extracted them into a separate bundle, reducing the load on the main site and eliminating duplicate JavaScript; Configured Webpack and the dependency structure: implemented tree-shaking and split dependencies between dependencies and dev Dependencies, improving the build process and fixing performance bugs; Ensured the stable operation of application forms in production: fixed submission errors, updated outdated links, and added missing UI features (breadcrumbs, card clicks, dark theme).