Nguyen Van Tinh — AI/MLOps Engineer

// About

Who I Am

I'm an AI engineer based in Vietnam who deploys AI agents into real engineering work — not demos. My strongest work lives where AI meets messy reality: automating a legacy-to-microservice migration across dev/prod, building the agentic engine of a Direct-to-Consumer commerce platform, and designing repeatable multi-agent processes that ship production code.

I think in systems and processes, not one-off scripts: orchestration patterns (Processor + Verifier, registry-driven teams), anti-false-pass verification, quality gates, and human-in-the-loop governance. I built a full AI-OS in ~11 hours by orchestrating a 5-agent dev team under a self-authored process — 27k LOC, 1,089 tests, zero live bugs.

Background spans computer vision (SETA International) and cloud-native MLOps (GCP/AWS/Kubernetes). I contribute to open-source tools like OpenCode (174k★). Now focused on AI automation & solution architecture for engineering and operations teams.

📍

Location

Quang Binh, Vietnam

🎓

Education

IUH — Computer Science

📧

Email

tinh2kqb@gmail.com

📱

Phone

0889 129 664

🏆

Achievement

1st Prize — IUH Science Conference

// Experience

Where I've Worked

Aug 2025 — Present · On-site

AI Engineer — Core Agentic Layer

PSA · PlatformDTC ↗

Built the agentic engine of PlatformDTC — an AI platform that builds & runs a Direct-to-Consumer business: store/landing auto-builder, marketing, support, support-email, and cart-recovery agents
Designed multi-agent workflows on Claude Agent SDK + MCP, turning D2C operations into autonomous agent pipelines
Contributed frontend, backend services, and the admin orchestrator coordinating & observing the agents
Contributed fixes to OpenCode (open-source Claude Code alt, 174k★)

2025 — Present · Remote / Contract

AI Automation Engineer

Mobile advertising & social platform (~millions of users)

Designed, built & operated an AI-agent automation system for legacy → microservice migration & QA across dev/prod — 56 APIs processed, 52 with full test coverage
Architected a 2-agent Processor + Verifier orchestration with anti-false-pass verification (verify-by-call, 3-run/env) — human-in-loop caught 2 prod bugs the automated verifier passed
Automated test-case generation from source (trace servlet → branch-coverage matrix → Postman/Newman → live verification)
Engineered multi-env data-safety gates: runtime credential resolution, capture-restore, prod write-protection

Oct 2024 — Aug 2025

Independent AI / MLOps Engineer

Self-employed

Architected production AI platforms (RAG, multi-agent systems, cloud-native infra) — turning business requirements into scalable systems on GCP/AWS/Kubernetes
Built MLOps/AIOps workflows: multi-agent orchestration, automated retraining, lifecycle management
Infra & observability: Kubernetes, Terraform, ArgoCD, Prometheus/Grafana/ELK

July 2022 — May 2024

AI Engineer

SETA International · Ha Noi ↗

Built & optimized production computer-vision models
End-to-end ML pipelines (data collection → deployment)
Deployed & monitored production AI services

// Projects

What I've Built

🧬

↗

Life OS

2026 · solo via a 5-agent AI team

Personal AI operating system — project/finance/automation tracing OS built in ~11 hours across 15 sprints by orchestrating an AI dev team under a self-authored process. 27k LOC · 1,089 tests · 12 modules · 0 live bugs. Registry-driven module auto-discovery, FastAPI + Next.js.

Multi-AgentFastAPINext.jsRegistry PatternQuality GatesClaude Code

🛠️

↗

Claude Manager (CCM)

Control plane for Claude Code

Observability + configuration management on top of Claude Code sessions. Live session tracing over SSE, cost analytics, token-burn dashboards, and 6 config managers (Agents, Skills, MCP servers, Plugins, Status line, Hooks). Single-user, loopback-only.

BunHonoReactMCPSSEOpenAPI

🛒

↗

PlatformDTC

Aug 2025 — Present · PSA (core AI)

Agentic platform that builds & runs a Direct-to-Consumer business with AI — store/landing auto-builder, marketing, support, support-email, and cart-recovery agents. I built the agentic engine + contributed FE, BE services, and the admin orchestrator.

Claude Agent SDKMCPNext.jsFastAPIMulti-AgentD2C

👥

↗

DevCrew

Registry-driven AI team platform · ~110k LOC

AI-agent team orchestrator (Kanban + chat + real-time SSE). Registry pattern: a new specialist team = 1 registry entry + 1 MCP + 1 prompt, zero new code. Event-driven orchestrator on Claude Agent SDK.

Claude Agent SDKRegistrySSEMCPFastAPINext.js

🧪

↗

Groundwork

QA & governance for AI agents · ~194k LOC

"Jira for AI agents" — API test lifecycle (Postman + MCP), human-AI dual-path (human owns setup, AI generates test logic), scoped agent tokens, audit trail, and a PII scrubber (AES-GCM). Built for governed, accountable AI-driven QA.

MCPPostmanPostgreSQLAgent TokensQA AutomationAudit Trail

📨

↗

OutboundOS

Agent-first outbound pipeline · ~122k LOC

Multi-phase prospecting pipeline (research → personalize → QA → classify → draft) with phase-gated evals. Agent-first runtime: business logic in agent prompts, Python as a thin tool/storage layer — hot-swap subagent ↔ SDK with no call-site change.

Claude Agent SDKAgent-FirstEvalsPythonPipeline

🧠

LexiOps · Crawl2Insight

Cloud-native AIOps & data platforms

LangGraph multi-agent AIOps for Kubernetes (per-tool RBAC, continuous retraining) + Vietnamese Legal RAG; and an end-to-end job-data platform on GKE with Terraform, ArgoCD GitOps, Airflow, and full observability.

LangGraphKubernetesTerraformArgoCDAirflowRAG

// Live Demos