Node.js and TypeScript Backend Engineer building production-ready infrastructure tools and distributed systems. Maintained systems serving 10M+ users at Hoichoi (OTT platform), built AxioDB — an embedded NoSQL database engine with 2,000+ NPM downloads — and currently developing NexoralDNS, a high-performance DNS server achieving 8,050+ QPS with 0.00% packet loss.

Focus areas: Database internals, networking protocols (DNS, TCP/IP), distributed systems, Node.js runtime architecture, RESTful API design, and microservices. I build tools that solve real problems I've encountered — from dependency compilation nightmares to home lab DNS management. All projects organized under Nexoral for production-grade development workflows.

6+ years of coding (2 years in production) with JavaScript, TypeScript, and Node.js ecosystem. Built and maintained systems serving 10M+ users, handling 40+ concurrent video streams, and managing 200+ IoT devices in production.

Core Stack (since 2020): Node.js, TypeScript, JavaScript, Express.js, Fastify, NestJS, React.js, Next.js, Golang Databases & Messaging: MongoDB, Redis, Redis Streams, RabbitMQ, SQL, Apache Kafka, Database Internals Infrastructure & DevOps: Docker, AWS Lambda, Cloudflare Workers, Nginx, Linux, Git, GitHub Actions, CircleCI Specialized: Video Streaming (FFmpeg, RTSP), IoT (MQTT, WebSockets), Microservices Architecture, RESTful APIs, GraphQL AI & LLM Integration: OpenAI API, Anthropic Claude API, Gemini API, Prompt Engineering, Structured Output (Zod), RAG Pipelines, AI Agents, Tool Use Additional: Firebase

My focus is building production-grade AI features inside real backend systems:

• LLM API Integration: Shipping services on top of OpenAI, Anthropic Claude, and Gemini APIs with retry logic, rate-limit handling, cost tracking, and latency budgets baked in.
• Structured Output & Validation: Using Zod schemas to parse and validate LLM responses — enforcing type safety at the boundary between language models and application logic.
• RAG Pipelines: Implementing Retrieval-Augmented Generation — embedding pipelines, vector search, and context injection — to give LLMs access to private or real-time data.
• AI Agents & Tool Use: Building agentic workflows where models call APIs, query databases, and compose multi-step actions autonomously.
• Prompt Engineering: Managing prompt versioning, designing system prompts for deterministic outputs, and building eval harnesses to measure output quality.

High-performance Docker-based DNS server for Local Area Networks. Built from scratch with custom UDP packet parsing, Redis caching, and Change Streams. Features web-based management interface and Docker deployment. Built with Node.js and TypeScript using dgram for UDP/TCP socket handling, Fastify for the API layer, and Next.js for the management dashboard.

Why built: Ever edited /etc/hosts on five different machines just to test one local domain? Ever wanted your own DNS server for your home lab without exposing it to the internet? That's why this exists.

Problem solved: Eliminates the hassle of managing /etc/hosts files across multiple machines, provides network-wide custom domain resolution, and adds security filtering for home/office networks.

Performance: 8,050+ QPS throughput, 0.00% packet loss, 500 concurrent clients, Redis-backed caching, 9-worker cluster

Tech Stack: Node.js, TypeScript, Docker, dgram, Redis, MongoDB, Fastify, Next.js

Lightweight embedded NoSQL database engine for Node.js applications. Pure JavaScript alternative to SQLite with MongoDB-style queries, zero native dependencies, and built-in web GUI at localhost:27018. Uses tree-like file structure for fast retrieval and worker threads for parallel processing. Optimized for 10K–500K documents.

Why built: Started building an Electron app and needed local storage. Tried JSON files first — worked fine with 50 records, got painfully slow at 1K+. Switched to SQLite, spent 6 hours fighting node-gyp rebuild errors across Windows and Mac. Deployed to production, got native binding errors. That weekend, I built AxioDB — pure JavaScript, no native dependencies, works everywhere Node.js runs. 2,000+ downloads on NPM later, turns out I wasn't the only one tired of this struggle.

Problem solved: AxioDB gives you MongoDB-like queries (find, insert, update, delete, aggregate) without the hassle. Just npm install axiodb and you have a database — no mongod process, no native bindings, no cross-platform compilation nightmares. Perfect for Electron apps, CLI tools, small websites, and anywhere you need a lightweight database that just works.

Tech Stack: Node.js, TypeScript, Worker Threads, Filesystem APIs

Universal Linux package builder that converts standalone binaries into native package formats (.deb, .rpm, tar.gz). Automates the creation of reproducible packages with configurable metadata, installation scripts, and service files. Designed for CI pipelines and Linux software maintainers.

Why built: I know many friends who code in C, Golang, Rust — they build, run ./bin, and it works. But ever thought about how to ship your binary to real users who install packages with sudo dpkg -i? That gap between "it works on my machine" and actual distribution is why this exists.

Problem solved: Streamlines the packaging process for Linux software distribution, automating repetitive tasks like metadata generation and file layout management. Makes CI/CD packaging simple with one command.

In Production: Currently used in ContainDB's Linux version packager for automated CI/CD pipeline builds.

Tech Stack: Go, Linux packaging formats

CLI tool for automating containerized database management. Provides instant setup of MongoDB, Redis, MySQL, PostgreSQL, and MariaDB with one-click installation of management tools (phpMyAdmin, pgAdmin, RedisInsight). Features Docker network integration, data persistence, and Docker Compose export/import capabilities.

Why built: Ever faced "core dumped" errors while installing MongoDB on your local Linux machine? Ever spent three hours debugging Docker network configurations just to connect pgAdmin to PostgreSQL? This tool was born from that pain.

Problem solved: Turns database environment setup from a multi-hour debugging session into a single command. Solves version compatibility issues, complex Docker configurations, and provides consistent development databases across teams.

Tech Stack: Go, Docker, CLI

Additional infrastructure and tooling projects available at the Nexoral organization.