Reza Salmanian

Command Palette

Search for a command to run...

All projectsOperations · Procurement

Vendor & Inventory

Vendor Risk & Inventory Operations Platform

A vendor and inventory module that tracks stock levels, procurement workflows, and vendor risk together, so purchasing decisions account for both cost and reliability.

Timeline

  1. Phase 1

    Data model

    Unified inventory and vendor performance model

  2. Phase 2

    Procurement workflow

    Built request-to-receipt procurement state machine

  3. Phase 3

    Risk monitoring

    Added vendor risk scoring from internal signals

  4. Phase 4

    Reporting

    Shipped combined operations reporting

Executive Summary

A vendor and inventory platform unifying stock tracking, procurement workflows, vendor risk monitoring, and business reporting, so operations teams can make purchasing decisions with both inventory and vendor-reliability data in view at once.

Business Problem

Inventory systems and vendor management tools are usually separate, which means purchasing decisions get made without visibility into vendor reliability, and vendor risk gets assessed without visibility into current stock exposure. Operations teams needed one view that connects what's in stock to who supplies it and how reliable that supplier has been.

Project Goals

  • Track inventory levels and movement in real time across locations
  • Give procurement a workflow that reflects real approval and ordering policy
  • Monitor vendor risk (delivery reliability, pricing volatility) alongside inventory data, not separately
  • Provide operational reporting that connects stock, procurement, and vendor performance

Solution Overview

Inventory levels, procurement requests, and vendor performance data share a common data model, so a low-stock alert can be evaluated alongside the reliability of the vendors who supply that item. Procurement workflows route through configurable approval steps, and a reporting layer surfaces the combined view for operations leadership.

Architecture Decisions

  • Linked inventory records directly to vendor performance history rather than treating vendor data as a separate reference table
  • Built procurement as a workflow (request → approval → order → receipt) rather than a single-step purchase form
  • Designed risk monitoring as a scoring layer over vendor history (delivery timing, price changes) rather than a manual rating field

Screenshots

Inventory tracking dashboard

Illustrative — not an actual screen

Procurement workflow status

Illustrative — not an actual screen

Vendor risk scorecard

Illustrative — not an actual screen

Architecture Diagram

Vendor & inventory architecture

Shared inventory / vendor data model
Procurement workflow state machine
Vendor risk scoring layer
Operations reporting

Technical Challenges

  • Keeping inventory counts accurate across multiple locations and concurrent updates
  • Defining a vendor risk score that's meaningful without access to external credit or financial data sources
  • Balancing procurement workflow flexibility against the complexity of configuring it for a new customer

Engineering Decisions

  • Started vendor risk scoring from internal signals only (on-time delivery, price stability) rather than waiting on third-party data integrations, to ship value sooner
  • Modeled procurement as an explicit state machine so every request has a clear, auditable status at all times

My Responsibilities

  • Designed the shared inventory/vendor data model
  • Built the procurement workflow state machine
  • Directed the vendor risk scoring approach and operational reporting layer

Technology Stack

TypeScriptReactNext.jsNode.jsPostgreSQLGoogle Cloud PlatformCloud Functions

Results

  • Gave operations a single view connecting stock levels to vendor reliability for the first time
  • Reduced procurement cycle time with a clear, auditable workflow
  • Surfaced vendor risk signals that previously required manual review of order history

Lessons Learned

  • Connecting inventory and vendor data in one model, rather than integrating two separate systems, made the risk-scoring feature possible in the first place
  • Shipping a risk score from internal signals first, instead of waiting for external data sources, delivered most of the value much sooner

Next project

Card2Manage

Continue