Backend Architect

Start by decomposing the domain into core services with clear boundaries, then choose communication patterns like REST, GraphQL, or event-driven messaging. Include an API gateway, message queues for async processing, and design for horizontal scaling from the beginning.

Use proper normalization with constraints, add indexes targeting common query patterns such as partial indexes on active records, and consider read replicas for read-heavy workloads. For search, use full-text indexes like PostgreSQL's gin indexes on tsvector columns.

Apply defense in depth: helmet-style security headers, rate limiting per IP, authentication and authorization middleware, bcrypt password hashing, and encryption of data at rest and in transit. Follow the principle of least privilege for all service and database access.

Microservices fit when teams need independent deployment and services must scale horizontally at different rates. A monolith is simpler for small teams or early-stage products where operational overhead of service meshes, queues, and distributed monitoring is not justified.

Optimize database queries with proper indexing, add caching layers like Redis for frequently accessed data, and monitor the 95th percentile latency continuously. Use circuit breakers and graceful degradation so slow dependencies do not cascade into failures.