NAMND

00 / 06Origin

Contact

Status: Senior ML Engineer at Success Software Services

Nam ND buildsAI systems forproduction.

Senior AI engineer focused on MLOps. Four-plus years shipping computer vision, NLP and generative AI, from GPU serving to the pipelines and teams that keep it running.

Explore the system Contact
Scroll to explore

01SUBJECT PROFILE

I take AI from notebook to production.

ND
ID / NGUYEN DUY NAM

I build production AI across computer vision, NLP and generative AI, plus the infrastructure that keeps it reliable: GPU serving, Kubernetes, CI/CD and monitoring.

I'm a Senior Machine Learning Engineer at Success Software Services. Before that I spent four years at FPT Software, where I led the MLOps platform and teams of up to eight engineers.

Role
Senior ML Engineer
Focus
MLOps and AI-Ops
Experience
4+ years
Certified
CKA and CKAD

Experience

  1. Jul 2025 – now

    Senior Machine Learning Engineer

    Success Software Services

    Smart warehouse and retail vision, LLM sales assistant, MCP developer tools with the CTO.

  2. Jul 2021 – Jul 2025

    AI Engineer, MLOps lead

    FPT Software, Quy Nhon City

    Smart-driving perception, an MLOps platform for 5+ AI teams, call-center NLP and drawing OCR.

  3. Mar – Aug 2020

    Data Analyst Intern

    SNAP Research Labs, Ho Chi Minh City

    Data pipelines for multi-market stock analysis.

02SERVICE MATRIX

What I build

Models are the easy part. I build the systems around them so they keep working under real traffic.

  • Detection / OCR / Edge

    Computer vision in production

    Multi-store and multi-camera vision, OCR for technical drawings, and edge inference on Jetson with TensorRT.

  • LangGraph / RAG / LoRA

    LLM and agent systems

    Multi-turn assistants, retrieval pipelines, LoRA and QLoRA fine-tuning, and MCP servers for developer tools.

  • Kubeflow / MLflow / CI/CD

    MLOps platforms

    Training-to-deployment pipelines, automated retraining, and monitoring with Prometheus, Grafana and Loki.

  • Triton / Kafka / Kubernetes

    Scalable model serving

    GPU serving behind event-driven pipelines, on AWS EKS or on-prem Kubernetes.

03SYSTEM CAPABILITIES

Tools I run in production

ML frameworks

  • PyTorch
  • TensorFlow
  • Hugging Face
  • OpenCV
  • Scikit-learn

Generative AI

  • LangGraph
  • LangChain
  • RAG
  • LoRA / QLoRA
  • Llama
  • Mistral
  • Ollama

MLOps and serving

  • Kubeflow
  • MLflow
  • NVIDIA Triton
  • TensorRT
  • ONNX Runtime
  • Prometheus
  • Grafana

Infrastructure

  • Kubernetes
  • Docker
  • Helm
  • Terraform
  • AWS EKS
  • Kafka
  • RabbitMQ
  • Redis

Certifications

  • CKA, Linux Foundation
  • CKAD, Linux Foundation
  • ML Engineer, Udacity
  • ML DevOps, Udacity
  • Cloud DevOps, Udacity
  • Data Engineering with AWS, Udacity
  • Data Streaming, Udacity
  • Machine Learning, Stanford / Coursera

04PROJECT ARCHIVE

Systems in the field

  • 90%less manual verification

    AI Smart Warehouse

    Event-driven platform on Kafka and RabbitMQ with Triton GPU serving. A new product onboards from just three photos.

    • Kafka
    • Triton
    • Grafana

  • 92%detection accuracy

    Smart Retail Analytics

    Multi-store vision with automated labeling, active learning and A/B testing, live in 10+ locations.

    • Computer vision
    • Active learning
    • Kubernetes

  • 1000+inquiries a day

    AI Sales Assistant

    Multi-turn LangGraph chatbot with Redis memory that lifted customer engagement by 40%.

    • LangGraph
    • Redis
    • Next.js

  • 80 msinference on Jetson Orin

    Collision Prediction

    0.92 mAP on vehicles and 0.88 on pedestrians. Auto-labeling made annotation 86% more efficient.

    • PyTorch
    • YOLOv8
    • Jetson

  • 60%faster deployments

    Enterprise MLOps Platform

    One training-to-serving stack for 5+ AI teams, with debugging time down 30%.

    • Kubeflow
    • MLflow
    • Kubernetes

  • 94%character accuracy

    Engineering Drawing OCR

    OCR and segmentation for Japanese blueprints, cutting document processing time by 65%.

    • PaddleOCR
    • DeepLabV3+

PHYRA, side project in development

An AI that solves physics problems and walks students through every step.

  • React Native
  • Expo
  • Supabase

05SIGNAL READOUT

Results, measured

  • 4+Years in production AI
  • 11Projects delivered
  • 8Engineers led
  • 5+AI teams on my platform
Impact delivered
  • Less manual product verification90%
  • More efficient annotation86%
  • Faster call processing75%
  • Faster model deployment60%
  • Higher live-stream engagement45%

Awards

  • 2023MLOps Marathon5th nationwide
  • 2023Kaggle ICR, age-related conditionsSilver medal, 199 of 6,430
  • 2022IT Race, Ho Chi Minh Open University1st place

06OPEN CHANNEL

Let's put your models into production.

Hiring, consulting, or an AI system that has to survive real traffic? Open a channel.