Skip to content

14. Tech stack recommendation

PageCovers
14.1 Backend servicesSpring Boot, modules, libraries, API gateway
14.2 Data and storagePostgreSQL, object storage, search, warehouse, ETL
14.3 Workflow and rulesWorkflow engine, rule engine, event bus, queue
14.4 Infrastructure and deploymentCloud, K8s, IaC, CI/CD, observability
14.5 Security and complianceIAM, secrets, encryption, audit, DR / BCP
LayerChoice
LanguageJava 17 LTS (or 21)
FrameworkSpring Boot 3.x
DBPostgreSQL 15+ (managed: RDS / Aurora)
Object storageS3 (or Azure Blob / OCI Object Storage)
CacheRedis (ElastiCache)
Event busRabbitMQ (MVP) → Kafka / MSK (scale)
WorkflowCamunda 7 (MVP) → Temporal (scale)
Rule engineJSON Logic + decision tables (MVP) → Drools (scale)
SearchOpenSearch
WarehousePostgreSQL replica (MVP) → Snowflake or ClickHouse
ETLdbt; CDC via Debezium
BIMetabase (MVP) → Superset / Looker (scale)
CloudAWS Mumbai primary, Hyderabad DR
OrchestrationEKS (Kubernetes)
IaCTerraform
CI / CDGitHub Actions / GitLab CI / Jenkins
SecretsAWS Secrets Manager + KMS (or HashiCorp Vault)
ObservabilityDatadog or Grafana + Prometheus + Loki + Tempo
Identity (internal)Google Workspace / Microsoft Entra ID + Okta
Identity (borrower)Mobile OTP + Aadhaar
EncryptionTLS 1.2+ in transit; AES-256 at rest with KMS keys
ML platformFeast + MLflow (at year 2+)
FrontendReact + Vite + TypeScript
MobileReact Native or native (if mobile app needed)
Field-agent appReact Native (offline-first)
  • Java / Spring Boot: team familiarity, fintech industry standard, deep library ecosystem, strong ops tooling.
  • PostgreSQL: ACID for finance; rich extensions; logical replication for CDC.
  • Camunda / Temporal: long-running workflows are first-class.
  • JSON Logic + decision tables: rules editable by non-engineers; promoted to Drools when complexity demands.
  • Kafka / RabbitMQ: event-driven async without distributed-tx pain.
  • AWS: best India-region availability, broadest managed services.

See sub-pages for detail.