DollTop Commerce & POS Platform
A commerce platform, not just a storefront
Three coordinated applications for a live retail business: a Next.js storefront, a centralized Express/PostgreSQL API, and an offline-capable Tauri desktop POS.
Software Engineer & Solutions Architect
I design and build software systems end to end — from requirements and architecture to production-ready web, backend, and desktop applications.
Full-stack software engineer with a growing focus on system design and architecture decisions across web, backend, and desktop applications. I'm a software engineer who works across the full stack: React and Next.js on the frontend, Node.js/Express APIs and PostgreSQL on the backend, and Tauri for desktop applications.
Presented as engineering case studies — the problem, the system, and the decisions behind it.
A commerce platform, not just a storefront
Three coordinated applications for a live retail business: a Next.js storefront, a centralized Express/PostgreSQL API, and an offline-capable Tauri desktop POS.
AI-assisted inventory management with strict per-account isolation
A full-stack business application for asset management: a searchable, sortable dashboard, full CRUD workflows, analytics, and AI-generated asset descriptions, with every record scoped to the signed-in account.
Marketplace access control enforced at the database layer
A real estate marketplace with public listing browsing, authenticated owner dashboards, admin moderation, and Stripe-billed subscription plans, authorized end to end with PostgreSQL row-level security.
Focused technical breakdowns of how these systems are actually built — problem, diagram, technology choices, and trade-offs.
01
Understand the problem
Clarify what the business or user actually needs before proposing a solution.
02
Gather requirements
Turn the problem into concrete functional and non-functional requirements.
03
Design the architecture
Define components, data flow, and boundaries between systems.
04
Evaluate trade-offs
Weigh complexity, cost, and risk for each viable approach.
05
Select technologies
Choose tools deliberately, based on the requirements — not by default.
06
Design data & APIs
Model the database schema and the interfaces that expose it.
07
Implement
Build with typed, maintainable, reviewable code.
08
Test and validate
Check the system against the original requirements, not just that it runs.
09
Deploy
Ship to production with a clear, repeatable process.
10
Monitor and improve
Watch how the system behaves in the real world and iterate.
Technologies I actually use in the projects above — grouped by capability, not listed as a logo wall.
The core stack behind every shipped project.
How systems are structured, secured, and kept consistent.
Where these systems run and how they ship.
Relational modeling, ORMs, and managed database platforms.
Applying AI inside product workflows and using it in the engineering process itself.
Beyond the browser: installable and mobile clients with local data.
An honest breakdown — what's backed by shipped work, and what I'm actively developing.
Requirements analysis
SRS documents, backlogs, and sprint structures at Intelligile; functional/non-functional requirements behind each featured project.
System design
DollTop's three-application architecture with a single backend source of truth.
Database architecture
Relational schema design across PostgreSQL, Prisma, Drizzle, and Supabase RLS policies.
API design
REST APIs (Express, mobile integrations) and typed Server Actions.
Authentication & authorization
Three different models applied to three different trust boundaries — see Architecture.
Integration design
Gemini AI, Stripe billing, and ESC/POS hardware integrated into production workflows.
Cloud architecture at scale
Current deployments are single-region (Vercel/Neon/Supabase); multi-region and high-availability design is an active learning area.
Observability & monitoring
No dedicated monitoring/alerting stack has been built yet — next step for production systems like DollTop.
Caching strategy
No dedicated caching layer implemented yet; current systems rely on database and framework-level performance.
Cost optimization & capacity planning
Formal cost modeling hasn't been required at current scale, but is a deliberate growth area.
Disaster recovery & backup strategy
Managed database providers (Neon, Supabase) currently cover backups; formal DR planning is not yet in place.
Full-stack delivery work and structured software-planning experience, plus a Computer Science degree and certifications.
View experienceOpen to full-stack engineering and solutions-architecture-adjacent work. Reach out by email or send a message directly.
Contact