Back to work
01Automotive Inventory & Operations Platform

PartYard

A custom internal web application built for an automotive parts business to manage vehicles, parts inventory, inspections, storage, reservations, sales, expenses and operational reporting from one system.

Private production application — public access unavailable.
Role
Full-Stack Developer
Stack
Next.js · React · TypeScript · MongoDB
Focus
Private Production System
partyard · internal
PartYard — Dashboard overview with live operational data

Automotive Inventory & Operations Platform

  • Donor vehicle and VIN management
  • Inspection workflow across large parts catalogues
  • Typo-tolerant multilingual parts search
  • Reservations, sales and transaction history
  • Operating result and date-range analytics

Overview

PartYard replaces a set of disconnected spreadsheets and manual routines with one internal system. Vehicles enter as donors, get inspected, become inventory, then move through reservation, sale and reporting — all inside the same application.

The system is used for daily operational work, so the priority was reliability, fast data entry and interfaces that hold up on a phone in a warehouse as well as on a desktop in an office.

What I Built

Vehicle Management

  • Donor vehicle records with VIN and vehicle data
  • Reusable vehicle templates for repeating models
  • Structured model, trim and configuration data

Inventory & Inspections

  • Guided inspection workflow per vehicle
  • Missing, damaged and not-for-sale part states
  • Storage locations and inventory movement

Smart Search

  • Part search across English, Russian and Georgian
  • Aliases, OEM numbers and transliteration
  • Keyboard-layout mistakes and minor typos handled

Reservations & Sales

  • Part reservations with clear status
  • Sales flow with transaction history
  • MongoDB transactional operations where required

Analytics & Reporting

  • Dashboard with date-range analytics
  • Expenses and operating result calculations
  • Excel export via ExcelJS

Media & Storage

  • Image uploads with presigned AWS S3 URLs
  • Delivery through AWS CloudFront
  • Validation with Zod across API boundaries

Smart Search

Warehouse staff do not type part names the way a database stores them. Search had to absorb that reality instead of fighting it.

01Three input languages — English, Russian and Georgian — against one catalogue
02Aliases and colloquial part names mapped to canonical entries
03OEM number lookup alongside plain-language queries
04Transliteration between Latin, Cyrillic and Georgian scripts
05Wrong keyboard layout resolved instead of returning nothing
06Typo tolerance for near-miss spellings

Architecture

Frontend

Next.js, React, TypeScript, Tailwind CSS

Application layer

Next.js route handlers and server logic

Database

MongoDB with Mongoose schemas and transactions

Media storage

AWS S3 presigned uploads, CloudFront delivery

Validation

Zod schemas shared across boundaries

Deployment

Vercel

Secure HttpOnly cookie authentication

Gallery

Dashboard overview with operational totals
Vehicle catalogue
Donor vehicle detail and specifications
Inspection workflow
Multilingual parts search
Active reservations
Sales and operating-result overview
Sales analytics and category breakdown
Storage locations overview
Detailed storage location
Vehicle inventory and part status
Mobile dashboard
Mobile inspection workflow
Mobile parts list
Mobile vehicle detail

Project privacy

Because PartYard is a private operational system, production access and confidential business data are not publicly available. The screenshots shown here are sanitized.

Next project

Professional Training and Employment Center

Vocational Training, Certification & Verification Platform