Portfolio Case Studies: Enterprise Resource Planning & Operations

[!NOTE] Project details and data have been anonymized to respect confidentiality agreements.

---

Case Study: Real-time Inventory & Waste Management System (Confidential Project)

Overview

This project involved developing a mission-critical system for tracking inventory and operational waste across multiple high-volume locations. The goal was to eliminate manual reconciliation errors and provide real-time visibility into stock levels and financial leakage.

The Challenge

• Data Fragmentation: Inventory data was siloed across different branches, leading to inaccurate procurement decisions.
• Waste Visibility: Operational waste (mermas) was tracked on paper, making it impossible to identify patterns or hold teams accountable.
• Concurrency Issues: Multiple users updating stock levels simultaneously often led to race conditions and data corruption.

Technical Decisions & Implementation

• Real-time Synchronization: Leveraged Supabase Realtime to ensure that stock updates are propagated instantly across all connected clients, preventing over-ordering or stockouts.
• Robust Audit Trail: Implemented a comprehensive auditing system that records every change to inventory levels, including the user, timestamp, and reason (e.g., sale, waste, restock).
• Custom State Management: Developed specialized React hooks (e.g., useInventorySync) to manage complex optimistic updates and error recovery during network instability.

Data Flow

graph LR
    User((User)) --> UI[Next.js Frontend]
    UI --> Hooks[Custom Hooks / Services]
    Hooks --> DB[(Supabase PostgreSQL)]
    DB --> Realtime[Realtime Engine]
    Realtime --> UI
    DB --> Audit[Audit Logs]

---

Case Study: Scalable Recipe & Costing Engine (Confidential Project)

Overview

A sophisticated engine designed to calculate the precise cost of complex products composed of multiple sub-components (sub-recipes) and raw ingredients, adjusting dynamically to market price fluctuations.

The Challenge

• Recursive Complexity: Products often consist of sub-recipes, which in turn consist of other sub-recipes, requiring a recursive calculation approach.
• Unit Conversion: Ingredients are purchased in bulk (e.g., kilograms) but used in precise measurements (e.g., grams, teaspoons), necessitating a robust conversion utility.
• Performance: Calculating costs for thousands of products in real-time could be computationally expensive.

Technical Decisions & Implementation

• Recursive Costing Algorithm: Implemented a depth-first search (DFS) algorithm to traverse recipe hierarchies and aggregate costs from the bottom up.
• Memoization Strategy: Used useMemo and server-side caching to store calculated costs for sub-components, significantly reducing redundant computations.
• Centralized Unit Utility: Created a strictly typed unit conversion library to handle diverse measurement systems (Metric, Imperial, Culinary) with high precision.

Architecture Diagram

graph TD
    subgraph "Frontend (Next.js)"
        UI[Recipe Management UI]
        Costing[Costing Engine Hook]
    end

    subgraph "Backend (Supabase)"
        Auth[Authentication]
        DB[(PostgreSQL)]
        Storage[Image Storage]
    end

    subgraph "External Integrations"
        POS[POS API Sources]
    end

    UI --> Auth
    UI --> Costing
    Costing --> DB
    UI --> Storage
    DB <--> POS