case study
My own SaaS — cloud storage, backup & data infrastructure · 2020 – Present
Everabyte
Founder & Lead Engineer
A SaaS platform for cloud storage, online backup, and data infrastructure. I own the product, backend, and cloud architecture end to end — from the authentication service and the encryption layer to billing — with multi-provider, S3-compatible storage across multi-region infrastructure.
Explore Everabyte →The problem
Anyone who has tried to back up and archive data across devices and teams knows the failure modes: fragmented storage, opaque encryption, provider lock-in, and costs that grow silently.
The solution
Everabyte brings cloud storage, online backup, and long-term archiving into one product: user and organization management, an S3-compatible storage API, subscription payments, and desktop backup applications — on multi-region infrastructure.
Architecture
A multi-service SaaS: a Next.js / TypeScript frontend, Node.js / Fastify API services, Go services for encryption-heavy and performance-critical processing, a PostgreSQL data model, and S3-compatible storage across multiple providers and regions — deployed through CI/CD.
Responsibilities
- Product, backend & cloud architecture
- API design · authentication · encryption
- S3-compatible multi-provider storage
- Multi-region infrastructure
- Billing & PostgreSQL data model
- Desktop backup applications (Flutter / Dart)
- CI/CD & deployment
- Technical strategy & team
Key challenges
Encryption-heavy operations were the first bottleneck — moving them to dedicated Go services fixed throughput without rewriting the whole product.
Provider independence: an S3-compatible storage layer keeps Everabyte free from a single cloud vendor.
Four architecture generations (Laravel → Node.js → Go services → Next.js), each driven by benchmarks and concrete product needs.
Architecture evolution
Laravel → TypeScript / Node.js → Go services → Next.js → AI-assisted workflow
Progressive rewrites driven by benchmarks and concrete product needs — performance, maintainability, and encryption-heavy processing moved to Go.
Product screens
Click a screenshot to zoom
AI-assisted engineering
AI coding agents are integrated into the engineering workflow for codebase analysis, progressive monorepo refactoring, assisted test generation, technical documentation, migration preparation, API contract review, bug analysis, and CI/CD assistance. Generated changes are manually reviewed, tested, and validated against architectural, security, and production requirements before being merged.
Agents are also used for structured multi-file changes across the Go, Node.js, and Next.js services — preserving application-specific engineering rules and documentation — and to explore implementation alternatives or identify potential regressions. Final architectural decisions remain manually validated.
Outcome
In continuous development and production since 2020 — progressively rewritten from its Laravel origins into a multi-service Go and Node.js platform.
Stack
Go · Node.js / Fastify · TypeScript · Next.js · Flutter · PostgreSQL · S3-compatible APIs · Payment integrations · Docker · CI/CD