Rondinelly
Castelo Braga
Backend Architecture · Distributed Systems · Integrations
I build scalable backend systems, integration platforms, and event-driven architectures for high-volume SaaS products.
Background
About
I'm a Senior Software Engineer specializing in backend architecture, distributed systems, and integration platform engineering for high-volume SaaS products. My work centers on building reliable, observable systems that scale under real operational pressure.
At Turno, I lead backend initiatives involving partner PMS integrations, Kafka-based event pipelines, and microservice architecture. I own systems end-to-end — from design and spec through implementation, rollout, and production monitoring.
I believe in spec-driven development, operational observability as a first-class concern, and using AI tooling to amplify engineering productivity. I work closely with product, partnerships, and engineering teams to align on technical strategy and deliver measurable outcomes.
Tooling
Tech Stack
Frontend
Backend
Messaging & Distributed
Databases
Infrastructure
Observability
Integrations
What I Do
Core Strengths
Backend Architecture
Designing scalable, maintainable service architectures for high-volume SaaS environments.
Partner Integrations
Building reliable OAuth, webhook, and REST API integrations with external PMS platforms.
Event-driven Systems
Kafka-based pipelines for async workflows, booking sync, and distributed processing.
Performance Optimization
Profiling, identifying bottlenecks, and delivering measurable throughput improvements.
Observability
Dashboards, alerting, and tooling that make complex distributed systems understandable.
Technical Leadership
Driving architecture decisions, technical specs, and cross-team alignment.
AI-assisted Engineering
Leveraging AI tooling to accelerate debugging, documentation, and implementation workflows.
Spec-driven Development
Writing clear implementation specs that reduce ambiguity and accelerate execution.
Team Collaboration
Partnering across engineering, product, and partnerships to deliver integrated solutions.
Incident Management
Designing safeguards, runbooks, and post-mortems that prevent repeat failures.
Results
Engineering Impact
Integration Service Microservice Migration
Led the architectural migration of partner integrations into a dedicated microservice, improving isolation, deployability, and team ownership.
Kafka-based Booking Sync Pipelines
Designed and implemented event-driven workflows that process high volumes of booking and listing synchronization events reliably.
Operational Visibility Through Dashboards
Built observability dashboards that transformed opaque distributed system behavior into actionable engineering insights.
Reduced Debugging Complexity
Introduced structured monitoring and logging strategies that cut time-to-diagnosis for production incidents.
Feature Flag Rollout Infrastructure
Coordinated scalable rollout strategies using feature flags, enabling safe incremental releases across partner integrations.
AI-assisted Engineering Workflows
Improved team engineering velocity by integrating AI-assisted debugging and development tooling into daily workflows.
Cross-team Technical Alignment
Led technical discussions across engineering, product, and partnerships for integration platform initiatives.
Spec-driven Architecture Changes
Created implementation plans and technical specifications that accelerated complex architecture transitions.
Operational Reliability Safeguards
Designed incident prevention mechanisms and early-warning systems that improved platform stability.
Production Work
Systems in Production
Backend systems and integration infrastructure built and owned in production at Turno — a high-volume SaaS platform for vacation rental operations.
Integration Service Microservice
MicroserviceContext
Vacation rental platforms depend on real-time synchronization with dozens of external Property Management Systems (PMS). Previously, integration logic was scattered across the monolith, creating deployment coupling and ownership ambiguity.
Architecture
Extracted integration logic into a dedicated microservice with clear bounded context. Service owns all partner communication, OAuth flows, webhook ingestion, and event publication to downstream consumers via Kafka. Exposes internal APIs consumed by the core platform.
Responsibilities
- Designed service boundaries and data contracts
- Led migration from monolith to microservice incrementally
- Implemented OAuth 2.0 flows for partner authentication
- Built webhook ingestion and validation layer
- Defined Kafka topic structure for downstream consumers
Technologies
Impact
Isolated deployment, clear ownership, and a foundation for onboarding new PMS partners without touching core platform code.
Kafka-based Booking Sync Pipeline
Event PipelineContext
High-volume booking and listing events from multiple PMS partners needed to be processed reliably, with retry semantics, ordering guarantees, and dead-letter handling — without blocking synchronous API flows.
Architecture
Event-driven pipeline using Kafka as the message backbone. Partner events are ingested, normalized, and published to typed topics. Downstream consumers process booking state changes, listing updates, and availability syncs independently. Consumer groups enable parallelism without coordination overhead.
Responsibilities
- Designed topic schema and partitioning strategy
- Implemented producer and consumer abstractions
- Built retry logic and dead-letter queue handling
- Defined event normalization contracts across partners
- Created observability hooks for pipeline health monitoring
Technologies
Impact
Handles peak booking volumes without degrading API response times. Failures are isolated and retryable without manual intervention.
Partner Observability Dashboard
ObservabilityContext
Diagnosing integration failures across multiple PMS partners was slow and relied on ad-hoc log queries. Engineering and partnerships teams lacked a shared view of integration health.
Architecture
Grafana dashboards backed by metrics emitted from the Integration Service. Key signals: event lag, error rates by partner, webhook delivery success rates, OAuth token refresh failures. Separate views for engineering (technical detail) and partnerships (operational health).
Responsibilities
- Identified key operational metrics worth instrumenting
- Implemented metric emission at critical integration boundaries
- Designed dashboard layouts for both engineering and ops audiences
- Set up alerting thresholds for critical failure signals
Technologies
Impact
Reduced mean time to detection for integration failures. Partnerships team gained self-service visibility without engineering involvement.
Feature Flag Rollout Infrastructure
Deployment InfrastructureContext
Migrating live partner integrations carries risk — any regression affects active bookings. Deploying to all partners simultaneously was unsafe. A controlled rollout mechanism was needed.
Architecture
Feature flag system enabling per-partner, per-integration enabling of new code paths. Flags evaluated at runtime against partner identifiers. Rollout proceeds incrementally: internal → low-volume partners → high-volume partners. Flags are observable through the dashboard layer.
Responsibilities
- Designed flag evaluation strategy for integration contexts
- Implemented flag-aware routing in the Integration Service
- Coordinated rollout sequencing with partnerships team
- Created monitoring views for flag-specific error rates
Technologies
Impact
Zero-downtime migration path for integration changes. Issues caught at low-volume partners before affecting high-value accounts.
Integration Incident Prevention Safeguards
Reliability EngineeringContext
Recurring integration incidents shared common root causes: unexpected partner API changes, token expiry without retry, and silent failures in async processing. Reactive debugging was the norm.
Architecture
Multi-layer safeguard approach: (1) schema validation on partner payloads at ingestion, (2) circuit breaker patterns on outbound partner API calls, (3) token health checks with proactive refresh before expiry, (4) anomaly detection alerts on event volume drops. Each layer emits structured events for observability.
Responsibilities
- Identified recurring failure patterns from past incidents
- Designed and implemented payload schema validation layer
- Introduced circuit breaker patterns for partner API calls
- Built proactive token refresh mechanism
- Created anomaly detection thresholds for event volume
Technologies
Impact
Significant reduction in incident frequency. Most previously-manual interventions now handled automatically with clear observability.
Side Project
Featured Public Product
Próximo Livro
AI-assisted book discovery platform
An AI-powered book recommendation platform focused on personalized discovery, programmatic SEO, and affiliate monetization. Helps readers find what to read next based on preferences and curated recommendation flows.
Where I've Worked
Experience
Turno
Senior Software Engineer
Backend and platform engineer focused on partner integrations, distributed systems, Kafka-based event pipelines, observability, and operational reliability in a high-volume vacation rental SaaS environment.
How I Think
Engineering Philosophy
Build systems that scale operationally, not only technically
Reliability and observability are product features
Good integrations require both technical and partnership thinking
Documentation and specs accelerate execution
AI should amplify engineering productivity, not replace engineering judgment
Get in Touch
Contact
Open to interesting backend and platform engineering opportunities. Let's talk.
Open to scheduling a call — reach out via email.