Pedro Proni

A multi-tenant online casino platform served as a whitelabel: one architecture, many branded houses. A microservices ecosystem spanning casino, payments, transactions, missions, segmentation, webhooks and game aggregation.

• Multi-tenant core with per-brand theming, feature flags and isolated data boundaries
• Composable backoffice — dynamic popup builder, missions engine and segmentation rules operators configure without engineers
• 20+ services under Clean Architecture, DDD and Event-Driven Architecture for team autonomy

20+ services · multi-brand · operator-configurable product

A low-latency distribution service that fans real-time events out to thousands of connected clients — the backbone for live, reactive player experiences.

• WebSocket transport tuned for low latency and predictable back-pressure
• Event-driven pipeline decoupling producers from connected consumers
• Horizontal scale with graceful degradation under connection spikes

Low-latency delivery · designed for connection spikes

A brand-aware bridge that routes and normalizes data across third-party providers and internal services — the seam that lets many brands share one integration layer.

• Per-brand routing driven by environment configuration, not code forks
• Deploy topology on ECS with S3-backed environment resolution
• Normalization layer isolating the platform from provider-specific quirks

Multi-brand routing · one integration layer for many houses

A predicate-based segmentation engine that classifies players in real time from behavioral and engagement signals — the targeting core behind CRM campaigns and ML segmentation.

• Composable predicate rules evaluated against live player signals
• Feeds CRM dispatch and machine-learning segmentation pipelines
• Designed for auditability and anti-fraud predicates

Real-time targeting · powers CRM & ML segmentation

The streaming router that ingests platform events and dispatches them to the right consumers — tracking, CRM, analytics — as an event-driven backbone.

• Topic/route mapping decoupling emitters from an evolving set of consumers
• Throughput-first design for high-volume event ingestion
• Instrumented for observability with request tracing

High-throughput event routing · the EDA spine

Recommendation and player-value systems: data pipelines feeding model training and real-time inference to personalize which games each player sees next.

• Data pipelines from Athena to serving stores for inference
• Player-value scoring and activity windows driving recommendations
• Real-time inference surfaced through platform APIs

Personalized recommendations · pipelines to real-time inference