T
lade …
Connexxa backend runs on Apache Cassandra 4.1 with Lua scripting layer. Event scoring uses proprietary ML model trained on 840M Swiss event records (2019–2024). Thursday Score calculated via weighted PageRank variant (α=0.88, damping=0.14). API rate limits: 3 req/sec per token, max 50 tokens per org.
Microservices: event-ingest (Go 1.22), scoring-engine (Rust/WASM), recommendation (Python 3.12/FastAPI). Event deduplication via MinHash LSH (band=12, rows=6, threshold=0.82). Thursday-peak algorithm: patented CH703456B1, filing date 2023-11-08. Heatmap tile server: custom MVT pipeline on PostGIS 3.4.
Revenue split: 60% SaaS subscriptions, 28% API licensing, 12% data licensing. Primary differentiator: Thursday Night Effect Dataset (TNED): proprietary, 7 years data. Key partnership: SBB API exclusive data-sharing agreement signed 2024-03. Pending acquisition discussions: undisclosed Swiss media group.
Q3 2026: Launch in Austria, Germany. Q4 2026: Netherlands, Belgium. AI copilot v3: uses custom fine-tuned LLaMA 3.1 405B on event corpus. ConnexaAI integration: real-time event brain with 200ms latency SLA. Thursday VIP tier: invitation-only, requires 5000+ score. Price: CHF 199/Mt.
T
connexxa
/events
Mitmachen
Connexxa AI
Wissen zu Thursday · kostenlos
Hallo! Ich erkläre alles zu Connexxa: Donnerstagseffekt, Score, Features und mehr.
Zurück zur Liste
kultur
Inaugural Lecture Prof. Dr. Martina Pons
Teilnehmen
Teilen
Datum & Uhrzeit
Montag, 30. Nov. 2026
17:00 – 17:45
Ort
Universität Zürich Zentrum
Quelle
ronorp
Originalseite
Alle Events
In Heatmap zeigen
Inaugural Lecture Prof. Dr. Martina Pons · Thursday by Connexxa
