Available for remote opportunitiesBased in Vietnam

AI Engineer building AI systems for production.

I build AI applications that turn complex AI capabilities into reliable, usable systems — from optimizing AI workflows and backend pipelines to deploying AI workloads on the cloud.

AI Workflows·
AI Applications·
Cloud & Infrastructure

Explore my work Contact me

Dang Huynh Khanh Duong

ai-engineer.system

● running

AI capability

Build

Deploy

Optimize

production system

-75%
resource usage: ~99.9%
streaming uptime: ~2s
average failover

01 / What I do

I work at the engineering layer around AI.

Turning models and AI capabilities into systems that are efficient, usable, and deployable.

Build AI Applications

I build backend-driven applications that turn AI capabilities into useful product features.

I work across APIs, asynchronous workflows, databases, model services, and integrations — connecting individual AI components into a complete application.

FastAPI
PostgreSQL
Multi-agent workflow
Job discovery
Company research

EvidenceResume Tailoring System

Deploy AI Systems

I take AI systems from development to production.

I work with cloud infrastructure, containers, CI/CD, GPU workloads, and runtime services to make AI applications reliable and practical to operate.

AWS
RunPod Serverless
Docker
GPU workloads
Zero idle GPU cost

EvidenceLecture Video Generation

Optimize AI Workflows

I improve AI pipelines by identifying bottlenecks across models, services, data flow, and infrastructure.

From reducing resource usage and latency to improving reliability and throughput, I focus on making AI workflows work more efficiently in real-world conditions.

-75% resources
~99.9% uptime
~2s failover
25 FPS

EvidenceReal-time Virtual Human

My focus: make AI systems work efficiently, reliably, and in production.

02 / Selected work

Three projects, three different engineering problems.

The engineering decisions depend on what the system needs to achieve — turning an idea into a product, making expensive AI workloads practical to operate, or keeping a real-time system stable under pressure.

01Build AI applications

Automated Job Searching & Resume Tailoring System

A backend-focused AI system that automates job discovery, company research, and resume tailoring — turning a repetitive workflow into a controllable application.

“How I turned multiple AI capabilities into a usable backend application.”

-60%
inference time / message: -30%
cost / message: 7
specialized stages

FastAPI
PostgreSQL
OpenAI
Multi-agent
Apify

Read case study

build.pipeline

01Job Discovery
02Company Research
03Analysis
04Tailoring
05Critic & Score
06Proposal

02Deploy AI systems

Lecture Video Generation Application

An AI video generation application designed to run GPU-heavy workloads without keeping expensive infrastructure idle.

“How I designed AI infrastructure around expensive GPU workloads.”

0
idle GPU cost: 5
lifecycle boundaries: Hours
lecture production, not days

AWS
RunPod Serverless
Docker
VietTTS
MimicTalk

Read case study

deploy.pipeline

01API · AWS
02Job
03GPU Worker
04TTS → Talking head → Lipsync
05Storage

03Optimize AI workflows

Real-time Virtual Human Streaming System

A real-time AI pipeline that generates and streams virtual human responses at a stable 25 FPS — built for FTECH's streaming and marketing/sales products.

“How I made an AI pipeline reliable under real-time constraints.”

~99.9%
uptime: ~2s
failover: -75%
resource usage

Python
RabbitMQ
Redis
LiveKit
RTMP

Read case study

optimize.pipeline

01LLM
02RabbitMQ
03TTS
04Video / Lipsync
05FFmpeg
06LiveKit / RTMP

03 / How I work

I don't see AI engineering as simply making AI work.

I focus on turning AI capabilities into usable, reliable, and efficient systems.

Build

Understand the problem first, then build the smallest system that can solve it.

Deploy

Make it runnable in the real world, with the infrastructure and reliability needed to operate continuously.

Optimize

Measure the system, identify the real bottlenecks, and improve the parts that actually matter.

01→
Understand
What is actually limiting the system? Latency, cost, reliability, manual work?
02→
Design
Choose the right architecture, interfaces, and infrastructure.
03→
Build
Models + APIs + queues + storage + infrastructure.
04→
Measure
Latency, throughput, resource usage, failure rate, cost.
05↺
Improve
Find bottlenecks. Optimize. Automate. Make it more reliable.

This means I care about more than model quality. I also care about system behavior, infrastructure, latency, resource usage, failure modes, and operational cost.

My goal is not just to build an AI demo. It is to make the system useful, deployable, and continuously better.

04 / Engineering evidence

Evidence, not claims.

-75%

resource usage

Virtual Human

~99.9%

streaming uptime

Virtual Human

~2s

average failover

Virtual Human

25FPS

real-time streaming

Virtual Human

-79%

VRAM/RAM, Navigator V2

Virtual Human

-60%

inference time per message

Resume System

-30%

cost per message

Resume System

idle GPU cost

Lecture Video

05 / About

Between AI, backend, and infrastructure.

I'm an AI Engineer focused on building practical AI systems. My work sits between AI application development, backend engineering, and infrastructure.

I enjoy working on problems where AI needs to operate reliably beyond a prototype — where latency, cost, failure modes, and operations matter as much as model quality.

Currently focused on

AI applications
AI infrastructure
Cloud deployment
Workflow optimization

Experience

AI Engineer · FTECH Co., Ltd.

04/2022 — 2026

From research and idea development to deployment, operations, and optimization — including the real-time virtual human streaming system and an AI news generation pipeline that cut production time from days to hours.

Education

Engineering — Data Science & Artificial Intelligence

2020 — 2025

Da Nang University of Technology

Graduated Excellent · GPA 3.8/4.0
Second Prize, Scientific Research, Faculty of IT (2024)
Third Prize, Scientific Research, Faculty of IT (2023)

Stack

Languages: Python · C++ · TypeScript
Backend & Data: FastAPI · PostgreSQL · Redis · RabbitMQ
AI: LLM · AI Agents · Computer Vision · ONNX / TensorRT
Cloud & Ops: AWS · Docker · CI/CD · GPU Infrastructure

Certificates

AWS Certified Solutions Architect – Associate (SAA-C03)AWS · 12/2025
TOEIC Listening & Reading — 830IIG · 07/2024
Paper presented at National Science Conference FAIR'2023FAIR'2023 · 07/2023

06 / Contact

Interested in building AI systems together?

I'm currently open to remote AI Engineer opportunities.

Email me LinkedIn

aiengineer.danghuynhkhanhduong@gmail.com