Feature
Customer Storefront
Product catalog, product detail pages, cart flow, checkout, QR payment path, and account pages built with Next.js App Router and React 19.
Project 01 / 2026
Year
2026
Role
Lead Developer, Project Manager, Code Reviewer
Technologies
Next.js, Java, PostgreSQL, FastAPI, AI

About this project
Smart PC Store is a complete e-commerce system for PC components and gaming gear. The platform combines a Next.js storefront, a Java Jakarta EE backend, and a Python AI microservice for product advice and price prediction.
Overview
Smart PC Store was built as a production-style commerce platform for PC components, gaming gear, suppliers, stock movement, and order management. The system is split into focused services: a modern customer/admin frontend, an enterprise Java REST API, and an AI server that enriches the buying experience with market-aware recommendations and price forecasts.
My Role
Led the project direction, backend architecture, frontend integration, code review, and delivery planning. The work included defining API boundaries, shaping the admin workflows, implementing authentication flows, coordinating AI integration, and keeping the multi-repo system deployable and documented.
Key Features
06 Items
Feature
Product catalog, product detail pages, cart flow, checkout, QR payment path, and account pages built with Next.js App Router and React 19.
Feature
Admin panel for products, categories, orders, users, suppliers, stock imports, and operational analytics.
Feature
Access-token and refresh-token flow with Axios interceptors, silent refresh, and race-condition-safe token rotation.
Feature
Jakarta EE backend with layered controllers, services, DAOs, DTOs, JPA/Hibernate entities, RBAC, and PostgreSQL persistence.
Feature
FastAPI microservice that answers product questions, reads product data, and returns suggested products for the storefront chat widget.
Feature
Forecast endpoint that uses supplier quotation history and Markov-style prediction to power 7-day price trend charts.
Tech Stack
Frontend
Backend
AI Service
The Challenge
The project needed to support both customer shopping flows and internal inventory operations while also connecting a separate AI service to live product and supplier-pricing data. Keeping authentication, admin CRUD, checkout, analytics, and AI responses consistent across three repositories was the main complexity.
The Solution
Separated the system by responsibility: Next.js owns UX and client state, Jakarta EE owns commerce rules and persistence, and FastAPI owns AI chat plus forecasting. The frontend talks to stable domain services, the backend exposes role-aware REST endpoints, and the AI server periodically syncs product and price-history data for grounded recommendations.
Results
04 Metrics
Impact
3
Impact
6
Impact
5
Impact
7
Lessons Learned
01
Multi-repo products need clear ownership boundaries early; otherwise frontend, backend, and AI changes become tightly coupled.
02
Admin workflows are easier to maintain when every domain owns a narrow service, store, and type contract.
03
AI features work better as product infrastructure when they return structured data, not only natural-language answers.
Screenshots
Swipe or use the arrows.
Related Work
03 Cases

Contribution to Kiro, an agentic IDE and CLI for spec-driven development, hooks, powers, MCP integrations, and codebase-aware AI workflows.

SDG-focused landing page, learning hub, and Notion-powered blog platform for the Youth for Change initiative at UD-UFLS.

Movie browsing interface built with Next.js 15, Supabase, Tailwind CSS v4, semantic theme tokens, and accessible UI primitives.