AutoProv

UK automotive dealers deal with a data mess. Service histories across 44 manufacturer databases. HPI checks, finance records, DVLA data in separate silos. Mileage readings lost between trades. Compliance paperwork done by hand.

AutoProv fixes that. One platform connects everything. Service records, odometer inconsistencies, Experian and DVLA checks, compliant PDI reports, Distance Sale Packs, and AI-powered dispute responses under UK consumer law.

I built an ingestion layer that queries multiple automotive APIs simultaneously, normalises conflicting schemas, and merges fragmented data into a single coherent record. Think of it as a translator for the UK automotive data ecosystem.

Speed was the real challenge. Some endpoints are slow by design. I optimised the async pipeline, parallelised upstream requests, and added caching layers. What used to take tens of seconds now resolves in moments, even under heavy traffic.

The hardest technical problem in the project. AutoProv compiles massive, deeply nested datasets into publication-quality PDF reports without hitting serverless timeouts or leaking memory.

I built a custom PDF generation subsystem that structures multi-tiered data into clean, readable multi-page layouts. Handles dense relational trees that would choke standard rendering pipelines. Professional-grade reports dealers can send straight to customers and regulators.

Deployed a full Stripe billing architecture with monthly and yearly tiers. The killer feature is an admin dashboard that lets project managers change pricing and update SKUs without a single redeploy. Everything syncs live with Stripe via webhooks that track lifecycles, handle failures, manage trials, and keep the database in lockstep with Stripe’s state.

Added a referral and rewards programme with client and server-side tracking. When a referred user converts, reward hooks fire automatically. Discounts or credits go to the referrer. No manual intervention, no support tickets.

I inherited a prototype that needed structural work. Rewrote the core. Modularised the codebase, decoupled tangled components, cut the technical debt. A clean, production-hardened platform that scaled from zero to commercial success.