Farzon Aslonov

🔧 Key Microservices:

• API Gateway — unified entry point for all client and internal requests, with centralized authorization, logging, rate limiting, JWT validation, and tracing. Ensures secure, observable, and consistent communication between services.
• Authorization Service — core access system: authentication, employee roles and permissions (RBAC), stateful JWT with Redis, full audit logging, and access control at route and action levels.
• Collections Service — centralized dictionaries (statuses, currencies, regions, etc.).
• Client Service — manages client profile and contact data.
• Pledge Service — asset tracking and collateral evaluation for loan applications.
• Business Process Service — visual and logical stage constructor for loan processing.
• Financial Analysis Service — flexible reports and calculations based on application and client profile.
• Loan Service — manages credit applications, statuses, operations, and history.
• Report Service — generates eventual reports for applications and clients.
• Document Service — generates .docx/.xlsx/.pdf: decisions, financial analysis, credit bureau data, etc.
• Storage Service — image and file upload, access, and storage.
• Scoring Service — integration with external scoring systems and internal ML-based evaluation.
• Notification Service — email/SMS/push delivery using Kafka with prioritization logic.

🧠 Architectural Features:

• Microservice architecture built with Go (Chi, Fiber, GORM), Python, Kafka, PostgreSQL, Redis, MongoDB, gRPC, and REST API. Docker and Docker Compose for containerization, NGINX for load balancing, full observability stack: Prometheus, Grafana, Open Telemetry, Loki/Promtail.
• Event-Driven approach via Kafka: key event flows (notifications, status changes) are implemented. Architecture prepared for future expansion with domain event publishing, analytical data transfer to Data Lake, and reactive workflows.
• Service Discovery and Registry using Consul: services register and discover each other automatically.
• Unified Redis Bus (distributed cache) — provides distributed caching, inter-service event bus, TTL-based storage, pub/sub support, invalidation, and logical segmentation per service.
• Full Observability: centralized metrics, tracing, alerting, and logging.
• Unified logging system: request-id, trace-id, span-id, user context (user_id, role_id, etc.), and deep error tracing at all layers.
• Architecture prepared for Kubernetes migration and requires minimal changes to run in a cluster — ready for load growth, team expansion, or hybrid deployments.

📊 Architectural Principles:

• Clean Architecture: clear layer separation, framework independence, testability, low coupling.
• DDD-light: domain-oriented business logic, usecase layer, separation of core logic from infrastructure.
• Standardized APIs and contracts: DTOs, mappers, client abstractions, coding style, unified agreements.
• Consistent design patterns: DI, middleware, error handling, wrappers over external clients.
• gRPC interceptors: logging, metrics, timeouts, cancellation, retry, panic recovery, tracing.
• 12-Factor App principles: ENV-based config, stdout logging, stateless design, scalable and environment-isolated services.
• CI/CD via GitLab: automated build, linting, tests, Docker image build, Harbor push, versioned environments, deploy automation.
• Harbor: private Docker registry with proxy-cache, vulnerability scanning, access control, and artifact retention.

🎯 Outcomes:

• Unified digital front-platform built to cover full lifecycle of applications, clients, reporting, and documents.
• Accelerated time-to-market: new logic and products launched without manual steps.
• Automated business processes reduced employee load and removed routine work.
• Anti-fraud and validation mechanisms: automated validation and detection of suspicious activity.
• Improved transparency and control via full observability: metrics, tracing, alerts, logging.
• System handles high load, scales horizontally, and remains resilient under failure.
Flexible access control: multi-role support per user (Multi-RBAC), authorization with session and TTL token management, action audit, and access control at the level of API routes and operations.
• Microservices accelerated development: team independence, fast time-to-fix, and time-to-deploy.
• Replaced outsourced system with an in-house, fully controlled and bank-owned platform.
• Engineering excellence established: CI/CD, code review, documentation, unified logging and monitoring.

🧰 In Development / Planned:

• Transaction Service — online payments and acquiring module, including P2P, QR, and core banking integrations.
• Mobile Gateway — dedicated gateway with auth, rate-limiting, and secure routing for upcoming client mobile app.
• Analytics & BI Service — Dashboard — aggregates service data and builds ClickHouse/Metabase reports and dashboards.
• Data Lake Gateway — anonymized event gateway for streaming data (Kafka → Lake on Cassandra/Spark), enabling ML training and audits.