The learner is a fullstack engineer with 4+ years at PT Bank SMBC Indonesia (Jenius), a digital bank serving ~6M users. Evidence of real production design surface: built Kafka-based event processing (~821k events/day), API/SFTP integrations with validation and deduplication, reconciliation workflows, load testing (k6/JMeter), capacity planning, production incident debugging, and mentoring. Tech stack: Node.js/NestJS/TypeScript, React Native, Kafka, MongoDB, Redis, Docker/K8s.
Depth claim: comfortable building and shipping the pieces (queues, caches, integrations) but self-assesses that design fundamentals feel thin — confirmed my diagnosis that their systems run at low QPS (~10 events/sec avg) and never strained, so they've never been forced to reason about scale-driven trade-offs. They recognized the reconciliation/dedup work as distributed-consistency problems without having the vocabulary.
Implications: don't re-teach building blocks; anchor every fundamental to systems they've already shipped. Prioritize estimation math, consistency/idempotency by name, partitioning/replication, and failure design. Learner chose to start with a calibration exercise before any lesson.