Building shou.be: a multilingual, AI-powered events platform
How ABSS Analytics designed, built, and operates a production AI application end to end — from serverless architecture to live AI features across seven languages.
The challenge
Build a real-time events discovery platform for Belgium’s after-dark scene — concerts, parties, theatre, and nightlife — that surfaces the right events in seconds, in the visitor’s own language, near their location. It had to aggregate data from many sources, stay fresh, work fluently across seven languages, and rank well in search — all while running lean and reliably in production.
This is the kind of project that touches every layer of modern software: data ingestion, cloud architecture, applied AI, internationalization, SEO, performance, and observability. It became the clearest possible demonstration of ABSS Analytics’ shift from analytics advisory to building complete AI products.
The approach
Rather than assemble a large team or reach for heavyweight infrastructure, the platform was built solo on a pragmatic, cost-efficient, cloud-native stack — proving that a single experienced engineer with the right architecture can ship something genuinely ambitious.
Architecture & data
A serverless backend handles search and API traffic with automatic scaling and near-zero idle cost. An EU-hosted managed database keeps data resident in Europe for compliance. A dedicated ingestion worker continuously pulls and normalizes events from multiple sources, deduplicating and structuring them into a clean, queryable model. A distributed cache and rate limiter keep the platform fast and resilient under load.
Applied AI
AI is embedded directly into the product rather than bolted on. An automated translation pipeline localizes content into seven languages, and semantic search using vector embeddings lets visitors find relevant events by meaning, not just keywords — so a search returns the right results even when the wording doesn’t match exactly.
Multilingual reach & SEO
The platform was built for international discovery from day one: correct hreflang signals, structured data, locale-aware pages, and edge caching for fast global delivery. Server-rendered social images generate shareable, on-brand previews automatically for every view.
Reliability in production
Production monitoring and error tracking run across the distributed services, with scheduled jobs keeping data fresh on a predictable cadence. The result is a live platform that stays healthy, observable, and maintainable — not a prototype, but a real product in operation.
What was built
Serverless backend
Auto-scaling API and search layer with minimal idle cost.
EU-hosted database
Managed Postgres with European data residency.
Ingestion worker
Continuous multi-source event aggregation and deduplication.
AI translation pipeline
Automated localization across seven languages.
Semantic search
Vector embeddings for meaning-based discovery.
Distributed caching & limits
Fast, resilient responses under real traffic.
Server-rendered images
Automatic shareable social previews per view.
Monitoring & scheduling
Production observability and reliable refresh jobs.
“The value of building our own platform is simple: we don’t just advise on data, analytics, and AI — we architect and ship it. Every recommendation we make to clients is grounded in having done the work ourselves.”
Why it matters for our clients
shou.be is proof that ABSS Analytics’ expertise now spans the full journey — from SAP and analytics strategy all the way to designing, building, and operating live AI software. For clients, that means a partner who understands both the business context and the engineering reality of shipping AI products, and who can move fluidly between advising and building.
- End-to-end capability — strategy, architecture, and hands-on delivery
- Modern, cost-efficient cloud-native engineering
- Practical, embedded AI — not AI for its own sake
- Production discipline: monitoring, compliance, and reliability built in
See it live, or talk about your own project
Explore the platform we built, or get in touch to discuss how ABSS Analytics can help you go from data and analytics to a shipped AI product.