01Research fieldHanoi, Vietnam · 2026

Ideas · systems · real-world impact

Ngo Trung Kien, research field

Researching intelligent systems. Building them for the real world.

I work across wireless networking, agentic AI, and applied machine learning.

Explore my work

Editorial portrait artwork, generated cutout study

Editorial portrait study · generated artwork

Knowledge
in context

02 / Selected systems

Research for a more intelligent and human world

Systems that connect research, human responsibility, and real-world work across intelligent communications and agentic AI.

01→
ScienzaOS
The platform for academic life
Product · Systems design · TypeScript
02→
MedAI
Vietnamese traditional-medicine assistant
Applied ML · Knowledge retrieval · Image recognition
03→
ActantOS
Runtime safety for autonomous agents
Agentic AI · Policy enforcement · Evidence tracing

Case index

01ActantOSRuntime safety for autonomous agents2024 · Present
02MedAIVietnamese traditional-medicine assistant2024 · Present
03AGRIFARM-AIAutonomous precision farming2024 · Present
04Intelligent CommunicationsHybrid VLC/RF management2020 · 2023
05ScienzaOSThe platform for academic life2025 · Present

Fig. 01 · Runtime safety for autonomous agentsResearch & systems design

ActantOS

Open siteResearch2024 · Present

Fig. 01 · Runtime safety for autonomous agentsResearch & systems design

Runtime safety for multi-agent AI systems: constrain delegated authority, coordinate shared approvals, and trace cross-agent effects.

Problem: Agents act through tool calls, and delegated authority is hard to bound once several of them share a workflow.
Solution: An enforcement baseline that intercepts tool calls, evaluates policy and approvals, and records evidence.
Result: The public v1.2.0 page presents that enforcement baseline as shipped, while broader multi-agent guarantees stay stated as research hypotheses.

Agentic AIPolicy enforcementEvidence tracing

Visit live site ↗Research & systems design

Public-site context only. This page does not claim ActantOS is secure, compliant, production-ready, or proven.

MedAI

Open siteResearch2024 · Present

Fig. 02 · Vietnamese traditional-medicine assistantResearch lead

A Vietnamese traditional-medicine AI assistant for herb lookup, folk remedies, interactions, sources, and safety guidance.

Problem: Traditional-medicine knowledge is scattered across herbs, folk remedies, and interactions, with no fast way to check a safety boundary or a source.
Solution: A Vietnamese traditional-medicine assistant for herb lookup, symptom analysis, drug-interaction checks, source references, and medicinal-plant image recognition.
Result: The public About page reports 5,190+ herb records and identifies HSB, Vietnam National University, Hanoi, with a 2024-present research phase.

Applied MLKnowledge retrievalImage recognition

Visit live site ↗Research lead

MedAI is not a doctor or medical facility and must not be used for diagnosis, prescribing, or treatment. No clinical, outcome, or regulatory claim is made here.

AGRIFARM-AI

Development2024 · Present

Fig. 03 · Autonomous precision farmingResearcher

An autonomous agricultural management platform integrating hybrid VLC/RF networks, multi-agent reinforcement learning, and agentic AI for precision farming.

Multi-agent RLVLC/RFIoT sensing

Researcher

Problem: Agricultural IoT needs connectivity that holds across shifting field conditions, where no single network covers the ground reliably.
Solution: An autonomous management platform that integrates hybrid VLC/RF networks, multi-agent reinforcement learning, and agentic AI for precision farming.
Result: Crop-aware network optimization is connected with reinforcement learning for intelligent handover across agricultural IoT. No deployment or outcome is asserted.

No deployment, partner, outcome, or production-readiness claim is asserted.

Intelligent Communications

Research2020 · 2023

Fig. 04 · Hybrid VLC/RF managementPhD research

Intelligent management systems for hybrid Visible Light Communication and Radio Frequency networks, from theory to experimental systems.

Problem: Hybrid VLC/RF networks have to hand over between light and radio without dropping a session, and manage a feedback channel while doing it.
Solution: Intelligent management systems for hybrid VLC/RF networks, carried from theory through to experimental systems.
Result: Published experimental models for seamless VLC/WiFi handover and prioritized feedback-channel architectures, plus a feasibility study and prototype for smart retail.

VLC/RFNetwork managementExperimental testbeds

PhD research

Specific experimental outcomes and deployment claims are not inferred from visual mockups.

ScienzaOS

Open siteDevelopment2025 · Present

Fig. 05 · The platform for academic lifeFounder & builder

A platform connecting teaching, research, and university-industry collaboration through shared sources, projects, review, responsibility, rights, and reuse.

ProductSystems designTypeScript

Visit live site ↗Founder & builder

Problem: Academic work fragments across teaching, research, and university-industry collaboration, so sources, review, and rights rarely share one foundation.
Solution: A platform connecting teaching, research, and university-industry collaboration through shared sources, projects, review, responsibility, rights, and reuse.
Result: Research Agent is presented as the first offer, with Teaching Studio and Business Collaboration on the same foundation; access is handled by request.

Official-site context only. This page does not claim ScienzaOS is production-ready, certified, enterprise-deployed, secure, or proven to deliver institutional outcomes.

03 / Profile

Research is the method. Working systems are the point.

Editorial portrait artwork, generated study

Editorial portrait study · generated artworkHanoi, Vietnam

Affiliation: Hanoi School of Business and Management, Vietnam National University, Hanoi

I work where networking, machine learning, and autonomous systems meet. My PhD built an intelligent management system for hybrid Visible Light Communication and Radio Frequency networks; that work now feeds into agentic AI, IoT for agriculture, and systems that govern how autonomous agents act.

Alongside the research, I build the platforms that carry it: ActantOS for runtime safety, MedAI for traditional-medicine knowledge, and ScienzaOS for academic work. I believe a claim is only worth making once it can be demonstrated.

Focus areas

01
Agentic AI & runtime safety
Constraining delegated authority, coordinating shared approvals, and tracing cross-agent effects in autonomous multi-agent systems.
02
Intelligent communications
Visible Light Communication and hybrid VLC/RF or VLC/WiFi networks, from experimental testbeds to seamless handover.
03
Applied machine learning
Reinforcement learning, edge deployment, and predictive models for agriculture IoT, energy, and public-safety communications.
04
Quantum-safe systems
Quantum communications and quantum-safe IoT security for networks that must stay trustworthy under new threat models.

Experience

2025 · Present
Lecturer & Researcher
Hanoi School of Business and Management, VNU · Hanoi, Vietnam
2024 · 2025
Lecturer & Researcher
Posts and Telecommunications Institute of Technology · Hanoi, Vietnam
2020 · 2023
Researcher
ENLIGHT'EM, University of Palermo · Palermo, Italy
2022 · 2023
Visiting Researcher
Toshiba Research Europe Limited · Bristol, United Kingdom
2018 · 2019
Research Trainee
VTT Technical Research Centre of Finland · Oulu, Finland
2014 · 2016
Wireless Communication Engineer
Ascent Communication Technology · Hanoi, Vietnam

Education

2023
Ph.D. in Information and Communication Technologies
University of Palermo, Italy
2019
M.Sc. in Wireless Communication Engineering
University of Oulu, Finland
2014
B.E. in Electronics Telecommunication Engineering
Posts and Telecommunications Institute of Technology, Vietnam

Capabilities

Agentic AI & multi-agent systems
Reinforcement learning
Visible Light Communication
Hybrid VLC/RF networks
Edge & distributed computing
IoT architecture
Quantum-safe security
Machine learning systems
Cloud architecture (Azure)
Systems & product design
Technical writing
Teaching & curriculum

Toolkit

Languages

Python
TypeScript
C/C++
MATLAB
SQL

AI / ML

PyTorch
Reinforcement learning
Multi-agent RL
MLOps

Systems

Network simulation
Edge deployment
Azure
IoT stacks

Web / Product

Next.js
React
Three.js
Design systems

04 / Research

The record, in the open

Peer-reviewed work in intelligent communications, applied machine learning, and multi-agent systems. Awards are noted where the venue recognised the contribution.

Journal
2026
Data driven adaptive SVC control for 110 kV substations using machine learning and edge deployment validation
Discover ElectronicsJournal
Book chapter
2026
Smart and Sustainable Agriculture: Offline Reinforcement Learning for Intelligent Handover in Hybrid VLC/RF Agricultural IoT Networks
LNEEBook chapter
Conference
2026
Multi-Agent Systems for Nontraditional Financial Security: Architecture, mechanisms, stress testing, and policy implications
CITA 2026Conference
Conference
2026
Towards food sustainability and security: Crop-aware network optimization using multi-agent reinforcement learning in hybrid VLC/RF systems
FutureTech 2026Conference
Book chapter
2025
A Predictive Machine Learning Framework for Proactive Customer Satisfaction Management in Electricity
LNEEBook chapter
Conference
2025
Smart and Sustainable Agriculture: Offline Reinforcement Learning for Intelligent Handover in Hybrid VLC/RF Agricultural IoT Networks
CSA 2025Conference
Conference
2025
A Predictive Machine Learning Framework for Proactive Customer Satisfaction Management in Electricity Sector
ICDTDE 2025Conference
Conference
2024
AI-Enhanced VLC/RF Hybrid for Smart IoT: A Revolution
ATC 2024Conference
Conference
2023
A Novel Intelligent Management System Architecture for Hybrid VLC/RF Systems in Smart Retail Environments
★ Best Poster Award & 3rd Place SRC
ACM MobiCom 2023Conference
Conference
2023
Seamless Handover in Hybrid VLC and WiFi Networks: An Experimental Model
★ Best Presentation Award
ICCCAS 2023Conference
Conference
2022
VLC/WiFi Hybrid Architecture with Prioritized Feedback Channel
FITCE 2022Conference
Conference
2021
Leveraging EDCA for Feedback Channel in Hybrid VLC/WiFi Architecture
MedComNet 2021Conference
Conference
2018
Distributed LSA Controller for Public Safety Communications
★ Best Presentation Award
IEEE ICCC 2018Conference
Journal
2018
Critical Communications over Mobile Networks: Use Cases for 5G with Spectrum Sharing, Network Slicing and QoS Control
IEEE AccessJournal

Recognition

2023
Best Poster Award · ACM MobiCom
2023
3rd Place · ACM MobiCom Student Research Competition
2023
Best Presentation Award · ICCCAS
2018
Best Presentation Award · IEEE ICCC
2018
1st Prize · Kickstart UniOulu Finals

05 / Books

Writing I want people to be able to use

An imprint I founded, and twenty-three titles across agentic AI, secure systems, intelligent networks, and practical AI guides. Three are published; five are in production; fifteen are publisher-adapted drafts awaiting expert review.

Published · The Research Ledger Press

Published 202601 / 03
Engineering Secure Agentic Systems
An Undergraduate Introduction to Building, Constraining, and Governing Autonomous AI Agents
Published 202602 / 03
Runtime Safety for Multi-Agent AI Systems
Principles, Patterns, and Practices for Safe and Reliable Autonomous Collaboration
Published 202603 / 03
Governing AI Agents Before They Act
Policy, Control, and Assurance for Trustworthy Autonomous Systems

In production

05 titles · working shelf

In production
Multi-Agent Systems for Systemic Risk and Resilience
Governance, Stress Testing, and Adaptive Coordination
In production
Intelligent Hybrid VLC/RF Networks
Architectures, Resource Management, and Beyond-5G Applications
In production
AI Agents for Smart and Sustainable Agriculture
Autonomy, Sensing, and Decision Support for Resilient Farming
In production
Edge Computing and Distributed Systems for the Internet of Things
Architectures, Orchestration, and Real-World Implementations
In production
Service-Oriented Computing and Software Engineering
Principles, Patterns, and Real-World Applications

ExBooks · Prepared editions

A wider working library

15 titles · publisher-adapted draft · expert review required

Agentic AI library

05 titles

Prepared draftA01
From Prompt to Agency
How AI Agents Reason, Use Tools, Remember, and Act
Prepared draftA02
The Agent Builder's Blueprint
Design Patterns for Tools, Memory, Context, and Multi-Agent Systems
Prepared draftA03
The Agent Reliability Loop
Evals, Observability, Testing, and Production Operations
Prepared draftA04
Permission to Act
Identity, Security, Guardrails, and Recovery for AI Agents
Prepared draftA05
Leading the Agentic Shift
Strategy, Workflow Redesign, Governance, and Human Accountability

Practical AI guides

10 titles

Prepared draftN01
Small Teams, Smart Systems
A Practical Guide to AI Workflows for Owner-Led Businesses
Prepared draftN02
Managing the Machine Team
How to Lead People and AI Without Losing Accountability
Prepared draftN03
When Familiar Voices Lie
A Verification Protocol for Deepfakes, Voice Clones, and AI Scams
Prepared draftN04
The AI-Wise Family
Raising Capable, Creative, and Safe Kids in an Algorithmic World
Prepared draftN05
The Proof-of-Learning Classroom
Teaching With AI While Protecting Thinking, Privacy, and Trust
Prepared draftN06
Human Advantage Compounds
Build a Career That Gets Stronger as AI Improves
Prepared draftN07
Local First, AI Second
A Practical Guide to Private Models for Sensitive Work
Prepared draftN08
The AI Buyer's Field Guide
Choose, Contract, Test, and Govern AI Vendors
Prepared draftN09
Projects With Judgment
Lead AI-Enhanced Work from Idea to Measured Outcome
Prepared draftN10
The Creator's Provenance Playbook
Use AI Without Losing Your Voice, Rights, or Audience Trust