DemandSync

• 

• 

Inspiration

The restaurant industry loses $162 billion annually to food waste while simultaneously facing stockout issues that impact revenue and customer satisfaction. This paradox presented a clear opportunity. Shah assembled a talented team to address this challenge. I joined as the Database Engineer because I saw the potential to build a data-driven solution that could meaningfully reduce waste and improve operational efficiency for restaurants.

What it does

DemandSync is a comprehensive inventory management and forecasting platform designed for restaurant operations. The system provides:

Executive Dashboard: Real-time visibility into key performance indicators including inventory turnover, cost of goods sold (COGS), profit margins, and cash flow metrics

Purchase Order Management: Automated ordering recommendations based on historical consumption patterns, lead times, and predictive analytics

Centralized Product Catalog: Unified database of inventory items with supplier information, pricing history, and availability tracking

Demand Forecasting Engine: Statistical and machine learning models that analyze historical data to predict future inventory requirements

Risk Management: Proactive alerts for stockout risks, price volatility, and supplier performance degradation Financial Analytics: Working capital analysis, inventory valuation, and optimization tools for procurement timing and order quantities

How I built it

Core Responsibilities - Database Engineering & Data Analytics: Backend Infrastructure (FastAPI + PostgreSQL)

Designed and implemented a normalized relational database schema encompassing products, suppliers, orders, forecasts, and transaction history Developed RESTful API endpoints using FastAPI with proper separation of concerns and dependency injection Implemented SQL models with appropriate relationships, constraints, and cascade behaviors Built comprehensive data validation layers and exception handling for system reliability Configured asynchronous operations for improved concurrent request handling

Data Analytics Pipeline (Pandas + NumPy)

-Architected ETL processes to transform operational data into analytics-ready datasets -Developed time-series forecasting algorithms using statistical methods including moving averages, exponential ---smoothing, and ARIMA models -Implemented aggregation functions for calculating business metrics: inventory turnover ratios, forecast accuracy (MAPE/MAE), and supplier performance scores -Created financial calculation modules for COGS tracking, gross margin analysis, and cash flow projections

FinTech & Risk Management Features

-Built risk scoring algorithms to quantify inventory vulnerability and stockout probability -Developed price volatility detection systems with threshold-based alerting -Implemented working capital optimization models analyzing payment terms and order timing -Created supplier reliability metrics incorporating lead time variance and fulfillment accuracy

Cross-Functional Contributions: -After completing my primary deliverables ahead of schedule, I expanded support to adjacent teams: -API Team Collaboration, Frontend Team Support

API Team Collaboration Support -Designed RESTful router architecture and endpoint patterns -Implemented query optimization through strategic indexing and caching -Added pagination, filtering, and sorting for large datasets -Developed comprehensive OpenAPI documentation

Frontend Team Support -Defined key performance indicators: inventory turnover rate, stockout frequency, forecast accuracy (MAPE), and supplier lead time variance -Designed metric scoring systems with color-coded thresholds (red/yellow/green indicators) -Architected data visualization components: time-series charts, comparative bar charts, pie charts, and real-time gauges -Built reusable React components and integrated Chart.js/Recharts with backend data streams

Challenges I ran into

-Designing a complex relational schema that balanced normalization and performance -Achieving real-time dashboard updates without overloading the backend -Building accurate forecasts with limited historical data

Accomplishments that I'm proud of

-Delivered core database and analytics infrastructure ahead of schedule -Built a scalable backend and API architecture capable of handling enterprise workloads -Achieved over 85 percent forecast accuracy during validation -Helped the next generation of developers via mentorship, guiding younger engineers and students on database ----design, API architecture, and data visualization -Contributed across backend, API, and frontend layers, demonstrating full-stack versatility -Built features with measurable business impact, capable of reducing waste by 20–30 percent and improving margins by 5–10 percent

What I learned

-Database optimization techniques such as strategic indexing, execution plan analysis, and connection pooling -React state management patterns for real-time dashboards and interactive UI -Deeper understanding of restaurant finance including inventory valuation, working capital, and cash flow modeling

Professional Development: -Strengthened cross-functional communication with both technical and non-technical partners -Improved ability to mentor younger developers and guide them through complex architecture decisions -Learned to manage competing demands, prioritize effectively, and deliver features iteratively -Developed product-thinking skills by translating operations workflows into intuitive, valuable features -System Design Insights:Balanced architectural trade-offs between normalized and denormalized database structures -Designed systems capable of serving both real-time operational needs and historical analytics -Learned API design patterns that support flexible frontend consumption and future extensibility

What's next for DemandSync

Short Term -Integrate advanced ML forecasting models (scikit-learn; Prophet) -Add multi-location support with centralized procurement and location-level analytics -Develop a mobile manager app for reviewing inventory and approving orders remotely -Expand analytics to include cohort analysis and predictive maintenance alerts

Long Term Vision -Launch a B2B marketplace connecting restaurants with suppliers -Use blockchain for transparent, immutable supply chain tracking -Multi-tenant SaaS architecture with tenant isolation -Integrations with major POS and accounting systems

Built With

• app
• azure
• azure;
• docker;
• fastapi;
• javascript;
• microsoft
• postgresql;
• python;
• react;
• service;
• shadcn
• sql;
• sqlalchemy;
• tailwind;
• typescript;
• ui;
• vite;