Projects that blend product thinking, clean UI, and scalable engineering.

A full stack monorepo blogging platform with a modern reading experience on the frontend and a Python FastAPI backend powering content workflows, approvals, and data persistence.

• Organized as a monorepo, covering the frontend, backend, and deployment workflow in one codebase.
• Frontend is deployed on Vercel, while the FastAPI backend and PostgreSQL database are deployed on Render.
• Uses Python and FastAPI on the backend to manage blog operations, structured APIs, and production-ready data handling.
• Includes SMTP-powered email notifications so approval-related blog actions can trigger mail delivery as part of the publishing flow.
• Designed as a practical content product with responsive UI, clean navigation, and full stack deployment experience end to end.

A full stack developer portfolio built to showcase real projects while also functioning as a production app with backend support, contact workflows, and offline readiness.

• Uses Next.js on the frontend with Python and FastAPI on the backend to support a full stack architecture instead of a static-only site.
• Implements Resend-based email handling so contact form submissions can be forwarded reliably.
• Includes service worker support for offline capability and a smoother returning-user experience.
• Designed with a mobile-first layout to present projects, technical skills, and contact paths clearly across screen sizes.

A data-intensive desktop application designed for end-to-end machine learning workflows, combining complex state management, high-performance UI rendering, and Electron.js delivery, and complex data-science-oriented user journeys.

• Worked on a highly visual, data-heavy application covering the full ML lifecycle: dataset ingestion, data engineering, preprocessing, feature engineering, model training, evaluation, versioning, and automated documentation.
• Handled complex state management across deeply nested workflows and interdependent UI components, ensuring consistency and stability across numerous edge cases in data transformations and model configurations.
• Implemented multiple data ingestion pathways including drag-and-drop uploads, direct file imports, and integrations with external data sources like Snowflake, Databricks, and MySQL.
• Contributed to an advanced visual data flow builder enabling dataset joins, transformations, and extensibility via Python-based logic before model training.
• Designed and developed UI for model lifecycle operations including node splitting/merging, evaluation dashboards, confusion matrices, sampling pipelines, and target variable configurations.
• Optimized performance for large-scale datasets using list virtualization techniques to efficiently render thousands of feature variables without UI degradation.
• Leveraged Web Workers to offload heavy computations (e.g., DOM-to-image processing), ensuring a smooth, non-blocking user experience.
• Adopted modern optimization strategies including React Compiler-driven automatic memoization to reduce unnecessary re-renders and improve responsiveness in complex UI trees.
• Applied secure coding practices and defensive programming to handle unpredictable user inputs and data inconsistencies in a production-grade environment.
• Built within an Electron.js environment, gaining experience in desktop application architecture, process management, and bridging frontend performance with system-level capabilities.