Case study
Auftrag One
An AI platform for public procurement that brings together tender notices from over 200 German, Austrian and EU sources and matches them to company profiles.
- My role
- Senior software architect
- Year
- 2022 – Present
- Product
- auftragone.com
Challenge
Public tenders are scattered across many portals. Companies need to identify relevant opportunities and assess deadlines, eligibility requirements and award criteria without manually reviewing every notice.
What I built
I built an event-driven platform using Kotlin, Go and Kafka. It ingests notices, classifies them into 331 topics across 25 sectors with Gemini AI and scores their relevance against company profiles. The platform runs on GKE with CI/CD, auto-scaling and automated backups.
Result
The platform has analysed over 530,000 notices, adds around 8,500 each week and serves more than 500 active users.
Key capabilities
- AI topic classification and relevance scoring for each company profile.
- Automated document analysis for deadlines, eligibility requirements and award criteria.
- A procurement network of buyers and bidders, with multi-organization access and team management.
Sources and context
Product information is based on the linked product website. Last reviewed: .
Auftrag OneHave a similar project?
Architecture, full-stack delivery, cloud and AI — tell me what you're building.