A production-ready task management REST API built to showcase a full Clean Architecture implementation with a complete observability stack, not just CRUD endpoints.

What it does:

• JWT authentication with RSA-signed access/refresh tokens and full token rotation
• Role-based access control (Admin / VIP / Standard)
• Task CRUD with status tracking (active, completed, closed) and full audit logging
• Per-endpoint rate limiting and Redis-backed session/user caching

What makes it interesting:

• Clean Architecture end-to-end — presentation, application, domain, and infrastructure layers stay properly decoupled, wired together with the Dishka DI container
• Full MELT observability: Prometheus metrics, Grafana dashboards, Sentry error tracking, and structured Loguru logging with rotation and JSON output for log aggregation
• CI/CD pipeline on GitHub Actions running lint (Ruff), type checking (Pyright), and a matrix test suite across Python 3.12–3.14 with real Postgres and Redis service containers, plus coverage gating (≥70%) uploaded to Codecov
• Fully containerized with Docker Compose, including optional profiles for PgAdmin, RedisInsight, and Grafana

Stack: FastAPI · SQLAlchemy 2.0 (async) · PostgreSQL · Redis · Alembic · Dishka · Pytest · Testcontainers

A digital banking REST API modeling real financial operations — multi-currency wallets, balances, and transfers — with the kind of data integrity and access control a banking system actually needs.

What it does:

• RSA-signed JWT auth (RS256) with access/refresh rotation and token revocation
• Debit and credit wallets protected by a hashed PIN, with block/unblock and soft-close flows
• Multi-currency balances (regular and foreign) per wallet, with admin freeze/unfreeze controls
• Deposits, withdrawals, and inter-wallet transfers with fee calculation
• Full transaction history with status tracking (pending / success / failed) and type classification
• Dedicated admin surface for managing users, wallets, balances, and transactions

What makes it interesting:

• Domain modeling that mirrors real banking constraints: User → Wallet → Balance → Transaction, where every financial operation is atomic and auditable
• Consistent modular layout per feature (api → contracts → service → repository) so every module is easy to navigate the same way
• Async Unit of Work pattern around every session/commit boundary — no partial writes on financial operations
• Deployed via Docker Compose with a dedicated migrations profile, keeping schema changes explicit and repeatable

Stack: FastAPI · SQLAlchemy 2.0 (async) · PostgreSQL · Alembic · Dishka · PyJWT (RSA) · Pydantic v2

_{Open to backend roles where reliability and clean architecture are part of the job, not an afterthought.}