I use AI as a core development tool while maintaining full control over architecture and code quality. My workflow combines strategic AI assistance with hands-on development expertise.
Development Environment & Tools
- Use Cursor AI as my primary IDE with integrated AI code completion, refactoring, and context-aware suggestions
- Apply custom code rules in Cursor to enforce consistent formatting, naming conventions, and project-specific coding standards
- Leverage GitHub Copilot for CI/CD workflows and automated testing integration
- Employ Claude AI as my main assistant for planning, problem-solving, and complex implementation guidance
- Actively discover emerging LLMs by monitoring Hugging Face, research papers, and developer communities for specialized models before they become mainstream
Advanced Prompting & MCP Integration
- Apply advanced prompt engineering frameworks (RACE, CARE, TAG) to get precise, high-quality outputs from LLMs
- Use Microsoft Copilot prompt generator to optimize AI interactions and extract maximum value from each query
- Connect Figma MCP server to access design elements directly in development environment for seamless design-to-code workflow
- Use Stripe MCP tools for embedding payment components and initializing Stripe setup with proper configuration
- Leverage language MCP tools for proper configuration, routing validation, and framework-specific code assistance
- Integrate specialized MCP servers from Hugging Face and Context7 for enhanced documentation access and development capabilities
Claude Skills & Custom Workflows
- Use MCP Builder skill for creating and optimizing Model Context Protocol servers
- Employ Web Artifacts Builder skill for creating complex multi-component web applications
- Utilize Canvas Design skill for generating visual assets and design elements programmatically
Strategic AI Usage
- Conduct detailed planning sessions with AI to develop comprehensive, step-by-step implementation strategies
- Use AI to identify edge cases, security vulnerabilities, and scenarios that traditional development might overlook
- Generate boilerplate code, API integrations, and complex algorithms with AI, then refine and optimize based on project requirements
- Leverage AI for code reviews to catch bugs and architectural issues before they reach production
Human-Led Development
- Write and architect all core application logic independently—AI accelerates implementation, not decision-making
- Critically review all AI-generated code to ensure it meets production standards and follows best practices
- Maintain deep technical knowledge to validate AI suggestions and catch potential issues
- Use AI as an intelligent pair programmer, not a replacement for engineering judgment
Business Impact
- Development Speed: Reduce feature development time through AI-assisted coding while maintaining code quality
- Cost Efficiency: Execute projects independently that would traditionally require either a full team or expensive third-party software by leveraging free, reputable LLMs that aren't yet mainstream.
- Storage Optimization: Use AI-powered image compression and format conversion (WebP/AVIF) to reduce storage costs and improve performance
- Rapid Prototyping: Build and validate MVPs in days instead of weeks with AI-generated descriptions, automated testing, and intelligent validation
- Proven Results: Transformed East Continental Gems from 0.02% conversion rate to over 900% increase through AI-assisted redesign, SEO improvements, and analytics integration
- Competitive Edge: Implement cutting-edge AI capabilities months before they become industry standard by discovering specialized LLMs early
This approach lets me move fast without sacrificing quality—combining AI speed with human expertise to deliver professional solutions efficiently.
Built a full-stack peer-to-peer vehicle rental platform (website and iOS web app) using SvelteKit, Svelte 5, TypeScript, Firebase, and Vercel. The platform enables users to host and rent exotic cars and yachts with integrated driver background checks, AI-powered image analysis and ID fraud detection, insurance coverage management, and role-based access control. Implemented AI-powered vehicle description generation, Google Places API for location services, real-time vehicle tracking, Stripe payment processing, and dedicated insurance portals with claim support. Developed separate CRM and Admin portals with multi-tier dashboards for super admin system oversight, admin operations, and user-level access restrictions.
I created a comprehensive analytics platform that enables users to subscribe and analyze sports trends with advanced insights. I implemented custom authentication, VPN blocking, and IP blacklisting to ensure secure and reliable access for subscribers.
For East Continental Gems—the only company certified to sell authentic gemstones that are certified and authenticated by Marvel—I inherited a struggling e-commerce site with a conversion rate of just 0.02%. I redesigned the entire site, implemented SEO improvements, and integrated analytics tools, resulting in over 900% increase in conversion rates.
Paw In One (In Development)
For Paw In One, I am currently developing a custom Shopify theme and a bespoke Shopify app that allows users to add and manage pets and sign up for pet subscription services. The app was built on Node.js and is hosted on Vercel. The website is in development and will launch soon.