Product Walkthrough
Short product walkthrough
Public menu and dish filtering
Customer order history
Admin order management
AI review summary
Overview
Full-stack system connecting ordering, operations, and post-delivery insights into a single, structured operational flow. Designed to coordinate customer, admin, and AI-driven processes within a single operational lifecycle.
Problem
The ordering and operations flow was fragmented across spreadsheets, chat, and disconnected tools, making it difficult to track orders reliably and increasing duplication and error risk.
- Menu, orders, and reviews were managed across separate tools with no shared structure
- Manual coordination across spreadsheets and chat led to duplication and inconsistencies
- Admins lacked a clear view of the full order lifecycle
Solution
We designed a unified, role-based product flow that connects customer ordering, admin management, and post-delivery insights into a single system.
- Centralized menu data, order states, and review insights into a single source of truth
- Structured the system around clear order states and separated customer and admin workflows to reduce complexity
- Introduced automated review summaries to streamline post-delivery analysis and reduce manual effort
Key Decisions
Separate customer and admin flows
Distinct interfaces reduce complexity and prevent conflicting responsibilities
Define explicit order states
Clear states improve tracking and predictability across the ordering system
Scope AI to post-delivery insights
AI is applied to review summarization without affecting the core ordering flow
Role
- Defined and refined end-to-end UX flows across customer and admin journeys
- Aligned UX decisions with business rules and API constraints
- Supported implementation and system consistency across the product
Process
Brief & Scope
Translated the product brief into a focused MVP, prioritizing core flows over feature breadth.
Flow Definition
Defined role-based journeys and access logic to structure customer and admin interactions.
Build
Worked across frontend and backend integration, connecting flows through REST APIs.
Rules & AI
Scoped the AI feature around a concrete operator need: summarizing customer reviews.
Delivery
Prepared repository, documentation, and demo to communicate system logic and decisions.
Architecture
The system is structured into distinct layers separating UI, state, API interfaces, and backend logic.
Application Architecture
Presentation
React 19, TypeScript, Vite, React Router, custom CSS system
State Management
React Context for auth and cart state, integrated with API calls via Axios
API Interface
REST endpoints for authentication, menu, orders, reviews, and AI summaries
Backend
Django 6, Django REST Framework, Simple JWT, service layer for business logic
Data
SQLite for local development, with PostgreSQL in production
Cloud Architecture (AWS)
Production architecture designed to scale the system, showing how frontend delivery, backend services, and data storage are distributed across AWS components.
Compute & Load Distribution
ECS Fargate behind an Application Load Balancer
Database
RDS PostgreSQL as the production database
Security
WAF, IAM, Secrets Manager, and private networking
Networking
Private backend subnets with controlled outbound access
Storage & Monitoring
S3 for assets and CloudWatch for logs and visibility
Challenges & Trade-offs
Flow separation vs system complexity
Separating customer and admin flows increases architectural complexity, but prevents responsibility overlap and keeps each interface focused and predictable.
State clarity vs flexibility
Defining explicit order states introduces stricter system rules, but ensures consistent tracking, predictable transitions, and clearer coordination across the system.
Outcomes
Outcomes
- Structured ordering, admin, and review flows into a unified system, reducing operational complexity across roles
- Defined explicit order states that improved tracking, predictability, and alignment between UX and implementation
- Introduced AI-driven review summaries to streamline post-delivery analysis without affecting core ordering
Key Learnings
- Complex systems require clear state definitions to remain scalable and predictable
- Separating roles and flows reduces cognitive load and prevents system conflicts
- AI is most effective when scoped to support workflows rather than replace core interactions
Tech Stack
Design & Delivery
Frontend
Backend
AI
Data & Cloud