From Raw Data to Report: Architectural Approaches in the Avtoteka Project
summary
Autoteka (autoteka.ru) collects vehicle history by querying over 10 external APIs plus static dealer data per VIN and timestamp, then normalizes disparate source schemas and deduplicates overlapping time records. The project uses VIN+event date as merge keys and applies rules-based merging to produce a consumer-ready event set; the talk describes the system architecture, how reports are generated, and why search and aggregation are implemented as separate microservices.
AI-written summary · about
chapters (14)
- 0:00Project introduction
- 3:38MVP development start
- 5:25Switching to RabbitMQ
- 7:14Encoding issues and testing
- 8:14Data source expansion
- 9:41Data validation checks
- 10:44Partner constraints and solutions
- 12:32Car collateral verification
- 14:21Payment and report delivery
- 16:08Current system architecture
- 17:33Microservices migration
- 19:39Security and constraints
- 21:05Results and achievements
- 22:30Q&A
related talks
28mTransforming Automotive Infotainment with Data
31mThe Road To Data Mesh: BlaBlaCar’s 3-year Journey To Self-serve Data Analytics
58mAvito Data Store: Analytics for a Microservices Architecture
29mAPI-fying the Engine: Replacing Legacy Automotive Protocols with REST and OAuth.
31mApidays Virtual Helsinki - Oliver Thamm, Xapix.io
37mAvito.AutoUpload: From 4M to 80M Active Listings
05mPivot
30mAn architecture for big data: Supporting journalism through transparency in advertising