Project Case Study

Recon Tool

A reconciliation platform that imports mixed-format datasets, normalizes records, and surfaces actionable mismatches through a web dashboard.

Status: Active Role: Full Stack Engineer Year: 2026 Type: Personal / Product Build
Live Project Source Code (private) Back to Projects

Project Overview

Recon Tool was built to solve a recurring operations issue: business-critical data arriving from different systems in inconsistent formats, with no reliable way to identify mismatches quickly.

The platform ingests files and feeds from multiple sources, standardizes records into a shared schema, then highlights variances, missing values, and conflict states in a decision-ready dashboard so teams can resolve exceptions faster.

Tools, Technologies, and Frameworks

Frontend: React / Vite. Dashboard UI, charts, filterable tables
Backend: Java / Spring Boot. API layer and reconciliation engine
Data: Supabase. Structured persistence + rule-driven transforms
Infra: Render (backend), Vercel (frontend). Hosted web deployment and environment config
Observability: Error tracking and operational logging
Testing: Validation of parser and reconciliation rules

Technical Deep Dive

Challenge

Input data varied heavily by source: inconsistent headers, partial records, conflicting identifier formats, and edge-case null handling that caused unreliable comparisons.

Solution

I introduced a multi-stage normalization pipeline with deterministic field mapping, schema validation, and rule-based reconciliation passes. This made comparisons stable, auditable, and easy to explain in the UI.

What This Shows About My Proficiency

  • Data modeling and transformation design under messy inputs
  • Building full-stack workflows from ingestion to analytics UI
  • Balancing product clarity with technical correctness
  • Designing for reliability, traceability, and iterative scaling