Anugrah Surya Putra

A Next.js app that generates professionally indifferent apologies. Perfect for when you need to say sorry, but don't really mean it.

Go-based REST API backend for a personal portfolio system using Clean Architecture, Gin, PostgreSQL, and SQLC. Architected a modular system using Clean (Hexagonal) Architecture to decouple business logic, database operations, and the HTTP presentation layer. Designed a LATERAL JOIN query using PostgreSQL JSON aggregation (jsonb_agg) to retrieve nested resume hierarchies in a single database roundtrip. Used SQLC for type-safe database queries and schema-first code generation, reducing boilerplate and improving performance with pgx/v5 connection pooling. Developed a pointer-based fetch-and-merge strategy to safely handle partial updates (PATCH/PUT) of domain resources without overwriting unset fields. Implemented table-driven unit tests using Testify and mock interfaces to validate use cases and HTTP endpoints.

Zero Fucks Given as a Service (ZFGaas) is a tiny API that returns random, unnecessarily verbose, and aggressively indifferent apologies for scenarios where you need to formally apologize for not caring.

Go-based REST API for retrieving Quranic texts, Tafsir (exegesis), and prayer times, featuring full-text search using Bleve and Redis caching. Architected a Clean Architecture backend in Go using the Gin framework, decoupling handlers, business services, and repository layers to ensure modularity. Engineered a high-performance full-text search engine using Bleve, configuring a custom N-gram text analyzer and Unicode tokenizer to enable fuzzy matches on transliterated queries. Optimized API response latency by integrating Redis as a cache layer with a 24-hour TTL for external upstream APIs, reducing downstream request overhead. Developed a custom IP-based fixed-window rate limiter middleware with Redis, using a fail-open design to maintain API availability during database or cache disruptions. Streamlined container deployments by building a multi-stage Docker pipeline that pre-builds and bundles search indexes to eliminate cold-start initialization latency.