Do Cao Hieu
I keep AI products running in production.
Ho Chi Minh City, Vietnam · self-taught in tech since 2016 · DevOps since 2023. No drama. No downtime theater.
AI-native infrastructure is a different discipline.
LLM inference costs scale with tokens, not requests.
Vector databases require tuning the way cache layers did a decade ago.
Agent orchestration is replacing cron jobs.
Now includes prompt regressions, evaluation drift, latency budgets, and model behavior.
I've been building and operating AI-native systems since 2025. The goal isn't deploying a demo — it's keeping production AI systems reliable, observable, and cost-efficient at scale.
From IT helpdesk to production AI.
// the camera flies the pipeline — each station ignites as you arrive
Automotive Engineering
Built a foundation in systems thinking, diagnostics, reliability, and problem solving.
IT & Technical Support
Set up and maintained the company's computers, network, and CNC machine systems. Became the de-facto technical person on site.
DevOps Engineer
Sharpened my English, then taught myself DevOps — moving from manufacturing systems to software systems.
AI Data · Outlier
Top 5% highest-performing contributor globally — remote AI training & evaluation, alongside contract technical work for Wellington Decorators.
DevOps & AI Infrastructure
Building and operating production AI workloads across a 5-server fleet — inference, observability, automation, reliability. Three products shipped and running.
The stack I run on.
Where I've operated.
DevOps Engineer & AI Infrastructure
- Design and operate 5 production servers running 97 Docker containers across a 5-node WireGuard full-mesh VPN.
- Full LGTM observability (Prometheus, Grafana, Loki, Tempo); CI/CD with Jenkins, Harbor registry, and Terrakube / Semaphore for IaC.
- Shipped and maintain IELTS Pocket, Clinic ERP and a self-hosted DevOps Lab Platform.
AI Data Contributor
- Top 5% highest-performing contributor globally.
- Train and evaluate frontier models (Claude, GPT, Gemini) through expert prompt engineering and code review.
IT & Technical Support
- Ran IT for 30+ staff — workstations, LAN / Wi-Fi, file server / NAS and backups.
- Maintained CNC machine systems and the production-floor data pipeline.
Things I've shipped.
Self-Hosted Infrastructure Fleet
5 production servers · 97 containers · one platform.
DevOps Lab Platform
Self-hosted interactive DevOps learning platform — browser-based, hands-on labs running in real containers.
DevOps Automation Toolkits
27 reusable IaC & automation repos — Ansible, Terraform / OpenTofu / Pulumi, ArgoCD / Tekton, CI/CD and monitoring stacks.
Engram-Mem
Persistent memory for AI agents — dual-memory (episodic vector + semantic graph) over CLI, HTTP, MCP & WebSocket.
IELTS Pocket
AI-powered IELTS preparation platform — automated Speaking & Writing grading, running in production.
Clinic ERP
ERP & dashboard for a physiotherapy clinic — scheduling, patient records, and billing. Built and running in production for a private client.
Floatwave
Floating mini-player for YouTube & YouTube Music on Windows — always-on-top, global hotkeys, queue, favorites, focus & cinema mode.
War stories, written down.
AI Agents in Production
A practical guide to building, deploying and operating AI agents in production — model selection, context engineering, cost management, observability and security. Based on real systems running Claude, GPT, Gemini and open-source models.
Get it on Gumroad →
Let's build reliable systems.
Whether it's AI infrastructure, platform engineering, observability, automation, or production operations — I'm always interested in challenging engineering problems.