AI · Automotive
A multilingual European platform combining price prediction, computer vision, real-time market analysis, and demand forecasting.
Role
Architecture, team lead, lead developer
Context
Automotive company based in Germany
Period
5 months of development
Platform
Web platform · TR / DE / EN
The challenge
Buying and selling used vehicles online depends on information that is hard to trust: is the price fair, does the listing match the car, how fast will it sell? Established marketplaces leave those questions to the user. The goal was a marketplace where AI answers them as part of the product—across several European markets and three languages.
The approach
The platform was designed as a set of microservices on Google Cloud, so that the AI components—pricing, vision, market analysis—could evolve independently from listings, search and user accounts. A team of three built it in five months, with the architecture, the core services and the AI integration led hands-on.
What was built
- 01
Price prediction engine
Estimates a vehicle's market value from its attributes and current market data.
- 02
Computer vision
Recognises and analyses vehicles from listing photos.
- 03
Market analysis & demand forecasting
Tracks the market in real time and forecasts demand for vehicle segments.
- 04
Multilingual by design
Turkish, German and English from the first release, serving users across Europe.
- 05
Microservice architecture
Node.js, TypeScript and NestJS services on GKE with Cloud SQL and Vertex AI.
The outcome
A complete, AI-first marketplace delivered by a small team in five months: pricing, image analysis and market intelligence are part of the core product rather than add-ons, and the architecture leaves room for each of them to grow.
Next project