Carrier Intelligence SaaS Platform

Freight brokers and logistics professionals needed a fast, reliable way to search millions of FMCSA-registered carriers to vet and connect with potential partners. The FMCSA portal is slow and hard to navigate, and existing third-party tools were expensive or limited in functionality.

The system needed to ingest millions of records from a large government CSV dataset, provide sub-second search results across DOT numbers, MC numbers, and company names, and support a full SaaS subscription model with tiered access control.

Designed and built the entire platform solo from zero. Created a streaming CSV import pipeline using Bun that downloads the FMCSA census dataset, normalizes records, and upserts in batches of 500 using ON CONFLICT resolution to handle updates without duplicating data. Built a ranked search API endpoint backed by custom PostgreSQL indexes.

The Next.js frontend includes carrier search with state, entity type, and authority age filters; detailed carrier profile pages with trust signal scoring; a saved carriers list; connection request flows; and a full SaaS dashboard with Stripe subscription billing across multiple pricing tiers. Supabase provides the hosted PostgreSQL layer and auth foundation.

The most interesting engineering challenge was the data pipeline. Streaming the CSV row by row rather than loading it all into memory was critical for handling a dataset this large on commodity hardware. The batch upsert strategy with ON CONFLICT needed careful tuning to balance write throughput against lock contention.

Building the full SaaS layer on top of the data layer reinforced how much work goes into the non-product parts of a product: billing, access control, dashboards, and onboarding flows all required as much care as the core search feature.