Overend.dev

Problem: The e-commerce landscape for print-on-demand products is saturated. Manually creating unique, high-quality designs and optimized listings across multiple marketplaces is time-consuming and unprofitable at scale.

Solution: I architected and built a fully automated, end-to-end pipeline. The system uses AI to generate unique artwork, programmatically creates compelling product mockups, and then pushes SEO-optimized listings to various marketplaces via their respective APIs.

• Architected a Rails backend to manage the entire workflow, from job queuing to final listing.
• Developed a Python/Flask microservice to handle the AI image generation and processing.
• Engineered a robust PostgreSQL database schema to track millions of unique designs, products, and listings.
• Integrated with Cloudflare R2 for asset storage and the Cloudflare API for domain management.

Outcome: This system allows a single operator to manage an entire e-commerce business that would typically require a team, creating a significant competitive advantage through radical efficiency.

Problem: Code reviews are essential but time-consuming. Understanding the intent behind large or complex `git diff` outputs can be a significant bottleneck for development teams.

Solution: I developed an open-source command-line tool that leverages AI to analyze `git diff` outputs. It generates clear, concise, human-readable explanations of the changes, helping developers quickly grasp the context of a commit.

• Built as a lightweight and flexible Python script, easily installable via pip.
• Designed to be model-agnostic, allowing users to plug in their preferred AI API provider.
• Created a simple Cloudflare Workers-based landing page for project visibility and documentation.

Outcome: WispGPT streamlines the code review process, saving developers time and reducing cognitive load. It has been well-received in the open-source community, demonstrating the power of AI to enhance developer workflows.

Problem: For testing, demonstration, or entertainment purposes, there was no simple way to control the outcome of Google's public-facing random number generator.

Solution: I developed a lightweight, open-source Chrome extension that injects a React-based UI onto Google's search results page. This UI allows the user to pre-determine the number that the random number generator will produce.

• Utilized the Chrome Extension API to safely inject and manage the UI on the target page.
• Built the user interface with React for a modern and responsive experience.
• Engineered the JavaScript logic to intercept and override the default RNG behavior without disrupting the rest of the page.

Outcome: The extension provides a simple and effective solution for a niche problem, showcasing skills in browser extension development, DOM manipulation, and front-end frameworks.

Problem: A major construction firm in Myanmar needed a comprehensive digital platform to showcase its properties, manage client relationships, and streamline its sales process in a rapidly growing market.

Solution: I led the development of a full-featured real estate platform. The system included a public-facing property showcase, a secure client portal for managing documents and payments, and an admin dashboard for sales pipeline management.

• Built on a robust Ruby on Rails backend, providing stability and scalability.
• Designed a PostgreSQL database to handle complex relationships between properties, clients, and transactions.
• Developed a responsive front-end with JavaScript to ensure a seamless experience across all devices.

Outcome: The platform successfully digitized the client's sales and marketing operations, improving lead generation, client management, and overall business efficiency.