Hoang Long Vi — Software Engineer

// 01 — about

Systems that stay fast under load.

I design and build backend systems that handle real concurrency — in Go, Python and Node.js — with a focus on performance, distributed design and advanced caching. The last stretch of my work has been deep in LLM-based systems: context optimization, multi-provider orchestration and production AI pipelines that real businesses run on. When a project needs it, I go full-stack — backend, frontend, real-time — and deliver the complete system.

6+

years in production engineering

e2e

architecture → code → deploy

24/7

commercially deployed AI systems

// 02 — skills

Toolbox

Backend & Systems

GoPythonNode.jsC# Distributed systemsHigh-concurrencyMicroservices & API design

AI Engineering & Agents

LangChainAI agents & tool callingn8n workflow automation LLM orchestrationRAG pipelines — structured & unstructured Context & token optimizationMulti-provider routing & fallback

Databases & Caching

Redis — advanced cachingPostgreSQLMySQLSQLite

Frontend & Client

ReactViteAngularUnity — real-time & perf

DevOps & Tools

GitHub Actions CI/CDLinuxDocker

// 03 — featured work

Things I've shipped

Biglead — AI Agent Platform

commercial · in production

Lead AI Engineer · Biglead · 01/2026 – present

LLM agents built with LangChain + tool calling that act on the CRM in real time — intent detection, automatic tagging
Context-optimization pipeline that cuts LLM token usage while keeping answer quality
Multi-provider routing & fallback — high availability even when a model provider degrades
Unified RAG over structured + large-scale unstructured data
Semantic routing layer dispatching tasks across multiple AI services
Agent & business workflows automated with n8n, wiring LLM services into CRM and messaging
Commercially deployed and actively serving businesses today

Node.jsPythonLangChainn8nPostgreSQLLLM APIs

Try it live ↓

We Craft — Multiplayer Game

ETechLab

Shipped a real-time multiplayer Unity game: low-latency networking, optimized message handling and Redis-backed state synchronization.

UnityGoRedis

▶ Google Play

Retail Chain POS & Ordering

contract · 4 months

Multi-store POS backend + customer-facing ordering interface, designed, built and delivered production-ready in four months — adopted by the client.

ReactNode.jsSQLite

Live demo ↗

// 04 — live demo

AI middleware, in action

This chat talks to the commercial AI middleware through its public streaming API — token-level SSE, straight from the agent. Pick a bot, watch the response stream in real time, and observe how the agent uses tools and retrieves information behind the scenes — every action is shown for transparency. Any external actions are simulated in a sandbox environment and are reported rather than executed.

Note: these demo bots are built around real-world Vietnamese use cases and are trained on Vietnamese knowledge bases. As a result, some domain-specific terms do not have direct or natural English equivalents.

The CRM tags shown on the right are streamed directly from each customer's live CRM. They preserve the original labels defined by the customer rather than translated versions, so what you see reflects the production data exactly.

// 05 — how it stacks up

Built for production, not just demonstrations.

Many AI chatbot platforms focus on visual workflow builders, which work well for a wide range of use cases. Our approach is different: we focus on an agentic AI backend designed for production environments, with native tool execution, retrieval, and automation built into the platform. Here's a high-level comparison:

Capability	Typical chatbot & workflow platforms	Our chatbot
Estimated infrastructure cost per message	~$0.012 – $0.024	$0.00046 – $0.0038
Agentic bot behavior + RAG	✕	✓
Auto-tagging & conversation classification	✕	✓
Sends images & large documents	✕	✓
Auto image recognition & search over the knowledge base	✕	✓
Custom tool calls, internal memory, external REST APIs	✕	✓

We're also happy to support deployments outside Vietnam. Feel free to get in touch if you'd like to learn more.

// 06 — experience

Track record

01/2026 — present

Lead AI Engineer · Biglead

Lead AI engineering for a commercially deployed AI agent platform
LLM agents with LangChain and tool calling, acting on the CRM in real time
Multi-provider routing & fallback, unified RAG (text + image), semantic routing
Context-optimization pipeline for token efficiency; n8n workflow automation

03/2023 — present

Online Multiplayer Game Backend Engineer & Game Developer · ETechLab (formerly EZTech)

High-concurrency backends in Go, Python and Node.js
Redis caching strategies for state sync and latency reduction
Shipped a real-time multiplayer Unity game in a small, fast team
Optimized network communication for low-latency gameplay

08/2021 — 03/2023

Online Multiplayer Game Full-Stack Developer · OHZE

Led development of a high-traffic online card game platform
Scalable backends in Scala, Python and C++
Runtime, memory and asset-pipeline optimization
Automated deployment; stable releases across WebGL, Android, iOS

10/2020 — 08/2021

Full-Stack Web Developer · FPT Software

Backend services in Java — Spring Boot, OSGi, Karaf
Redis caching to cut database load
Responsive Angular frontends for an enterprise platform

06/2020 — 09/2020

Backend Developer Intern · LinxHQ

Backend features for web applications in PHP with the Yii2 framework

2018 — 2021

Higher Diploma in Software Engineering · BachKhoa-Aptech International IT Training Academy

Graduated with Distinction · Software Engineering · Data Structures & Algorithms

// 07 — contact

Let's build something production-ready.

Open to remote and contract opportunities in the Netherlands, as well as international collaborations on AI systems, scalable backends, and commercial deployments.

lynxnymphy@gmail.com Schedule a call

📍 Hanoi, Vietnam GitHub