The brief
Build and scale a comprehensive AI-powered automotive retail platform enabling car dealers and OEMs to manage inventory photography (AI backgrounds, 360 spin, video tours), publish branded dealer websites, execute digital contracts, manage leads via CRM, and gain real-time analytics — all from a unified console with white-label mobile SDK capabilities.
What we built
A full-stack automotive AI SaaS platform built across 13 interconnected systems over 94 weeks with 15 contributors:
- A HubSpot-integrated digital contracting module with DocuSign/Signeasy e-signatures, versioned flows, and role-based access
- A white-label dealer website builder with drag-and-drop CMS, multi-theme, SEO tools, plugin marketplace, and multilingual support
- An AI image processing studio with background replacement, 360-spin, number plate blur, and classification across web and native apps
- A native iOS SDK (SpyneSDK using RealmSwift/SPM) and Android app with offline-first inventory, parallel uploads, and enterprise switching
- Real-time analytics dashboards powered by a Debezium→Kafka→ClickHouse CDC pipeline
- A video tour/360 module with voice-over, templates, and reel-mode pages
- A plugin marketplace with Google Vehicle Listing, Facebook Marketplace, and social feed integrations
- A full CRM with call recording (AES-256), customer cards, KPI analytics, and AI conversation metrics
- An AWS infrastructure migration from AMD→Graviton, g5→g6 GPU, self-managed Kafka→MSK, and observability stack modernization
- A website scraping pipeline (Webmine) with 85+ dealer sites, Kafka consumers, AWS Step Functions/Lambda.



