Hello, I'm

Nguyen Ngoc Hai

AI Agent / LLM Application Engineer

Nguyen Ngoc Hai

About Me

I am an AI Agent / LLM Application Engineer with a Bachelor's degree in ICT from the University of Science and Technology of Hanoi.

I build LLM-powered applications, RAG workflows, and AI agent features with Python, FastAPI, LangChain, Langfuse, DeepSeek and OpenAI APIs, and PostgreSQL.

My work uses retrieved context, source citations, structured outputs, and constraint-based planning to support dependable AI workflows—from backend integration to cloud-native deployment and observability.

🎓

University of Science and Technology of Hanoi

Bachelor of Information Communication and Technology

Oct 2020 – Oct 2024

VinUniversity (VinUni)

AI in Action

May 2026 – Aug 2026

AI Agent Engineering

Building LLM-powered products with LangChain, DeepSeek and OpenAI APIs, RAG, AI agents, and structured outputs.

Grounded RAG Systems

Designing retrieval pipelines and PostgreSQL/pgvector-backed data access for context-grounded answers with source citations.

LLMOps & Observability

Tracing, monitoring, and debugging AI workflows using Langfuse, OpenTelemetry, and Grafana tooling.

Cloud-Native Delivery

Deploying AI applications with Docker, Kubernetes, Terraform, Ansible, Helm, observability, and CI/CD.

Backend & APIs

Developing reliable FastAPI and Node.js services with REST APIs, WebSockets, PostgreSQL, and async processing.

Reliable AI Workflows

Applying structured outputs and constraint-based planning to make AI-generated workflows more dependable.

Experience

AI Application / Full-stack Developer

Pancake

Jul 2025 – May 2026

Hanoi, Vietnam

  • Built an LLM-powered customer support agent using Python, FastAPI, LangChain, retrieval-augmented generation (RAG), and WebSocket streaming.
  • Designed context retrieval, prompt construction, conversation handling, and structured-output workflows to generate relevant responses from product and support data.
  • Integrated Langfuse for LLM tracing, prompt and response monitoring, latency analysis, and debugging of retrieval and generation workflows.
  • Developed and maintained Python and Node.js backend modules with PostgreSQL-backed data access, REST APIs, and asynchronous processing, focusing on reliability and maintainability.

Software Engineer

Ekotek Software and Blockchain Development Company

Mar 2024 – Jul 2025

Hanoi, Vietnam

  • Progressed from a three-month software engineering internship to a full-time fresher role, contributing to both frontend and backend development.
  • Built responsive web applications using TypeScript, React, Next.js, Tailwind CSS, Ant Design, and Material UI.
  • Managed complex client-side state and server data synchronization using Redux Toolkit and TanStack Query.
  • Developed and integrated Node.js backend modules, REST APIs, PostgreSQL, Prisma ORM, and JWT-based authentication workflows.
  • Collaborated through Git-based development, code reviews, API integration, Docker-based environments, and CI/CD workflows.

Selected Projects

Cloud-Native RAG Platform

KubeRAG

Aug 2026

A self-hosted RAG platform for Vietnamese news, designed for cited answers and production-grade cloud-native operation.

  • Built and deployed a platform that indexed more than 1,000 Vietnamese news articles and generated answers with source citations using PostgreSQL/pgvector and llama.cpp.
  • Deployed on a three-node k3s cluster using Docker, Terraform, Ansible, Helm, and Kustomize.
  • Implemented PostgreSQL replication, Envoy rate limiting, and OpenTelemetry/Grafana observability.
FastAPIPostgreSQLpgvectorPrefectllama.cppKubernetes/k3sGoogle Cloud PlatformTerraformOpenTelemetryGrafana
LLM + Optimization

AI Academic Advisor

June 2026

An AI academic advising application that generates semester-by-semester study plans from curriculum rules, course offerings, and student records.

  • Implemented a Google OR-Tools CP-SAT engine to satisfy prerequisites, corequisites, course availability, credit limits, and graduation requirements.
  • Designed FastAPI APIs and a PostgreSQL/Supabase schema for academic planning data.
  • Integrated an LLM layer to explain study plans and academic warnings in student-friendly language.
FastAPIPostgreSQLPythonOR-ToolsAI AgentsREST APIs
AI Agent + Full-stack

FinTrack

Jan 2026

An AI-powered personal finance application that converts Vietnamese natural-language inputs into structured financial transactions.

  • Implemented intent recognition, entity extraction, and semantic mapping to resolve transaction amounts, types, categories, and wallets.
  • Built the application with a NestJS, Next.js, Prisma ORM, FastAPI, and PostgreSQL stack.
  • Mapped natural-language transactions to reliable PostgreSQL entities for personal finance tracking.
NestJSNext.jsPrismaFastAPIPostgreSQLTypeScriptPython

Skills & Technologies

AI & LLM

LangChainLangfuseLangGraphDeepSeek APIOpenAI APIllama.cppRAGAI AgentsEmbeddingsSemantic SearchPrompt EngineeringStructured Outputs

LLMOps & Observability

PrefectOpenTelemetryPrometheusGrafanaLokiTempoPyroscope

DevOps & Cloud

DockerKubernetes/k3sTerraformAnsibleHelmKustomizeGitHub ActionsCI/CDGoogle Cloud PlatformLinuxGit

Backend & Data

PythonFastAPINode.jsNestJSExpress.jsPostgreSQLpgvectorPrismaREST APIsWebSocketsJWTSQL

Frontend

TypeScriptReactNext.jsTailwind CSSRedux ToolkitTanStack QueryAnt DesignMaterial UI

AI Quality Focus

Context-grounded retrieval, source citations, structured outputs, and constraint-based planning for dependable AI workflows.

🌐

Languages

EnglishB2 — Professional
🏆

Certifications

Introduction to Programming and Problem-Solvingby Big-O • June 2023View credential

Get In Touch

I'm currently open to new opportunities. Whether you have a question or just want to say hi, I'll try my best to get back to you!

📍

Location

Hanoi, Vietnam

Open for both remote and on-site opportunities. Feel free to reach out!

Send me an email