AIVN360Get in touch

AI-native engineering & education

Engineering expertise.
Computation you can inspect.

AIVN360 builds engineering software and learning experiences that connect domain expertise, AI agents and deterministic computation.

Starting with railway-bridge dynamics. Built for engineers who need to understand the result—and the work behind it.

Explore RailDyn Talk with us
RAILDYN / ENGINEERING NOTE 001In development
Illustrative structural model of a train crossing a simply supported railway bridge, with moving axle loads and a span dimension.
Moving loads. Structural response.Concept illustration · not simulation output

Engineering decisions need more than an answer.

They need explicit assumptions, dependable calculations and evidence that can be reviewed. That is the foundation of our work.

01 / Our approach

Intelligence around the work.
Rigor at its core.

We are developing workflows in which AI assists with the surrounding tasks, engineering solvers perform the calculations, and people retain judgment.

01

Start with the engineering

Models, assumptions and boundary conditions come first. Domain knowledge gives the software a meaningful problem to solve.

02

Calculate deterministically

Numerical solvers handle engineering computation. AI assistance does not replace the mathematical model or the checks it requires.

03

Make the evidence visible

Inputs, units, model choices and verification results belong alongside the output, so the work can be understood and reviewed.

02 / Flagship product

RailDyn

Understanding the bridge.
As the train passes.

Active development

Railway-bridge dynamics software, grounded in numerical engineering.

KD-Railway, the Rust engine behind RailDyn, analyses structural response to moving train loads. The work includes moving-load and vehicle–bridge interaction calculations, with structured outputs for engineering workflows.

The development focus is a traceable path from model assumptions to engineering results, with agent-assisted workflows as the next direction.

See the reference case
RECORDED REGRESSION EXAMPLEHSLM-A1 / 20 m span

One model. Three train speeds.

Peak displacement at three sampled speeds from a recorded software regression fixture. Values are recorded, not re-run for this website; bars use a 0–8 mm scale.

Model, results & provenance

RailDyn uses KD-Railway, a Rust computation engine that builds on CALDINTAV (see attribution). The product is under development; software verification does not establish validation for engineering design use.

03 / The company

Built from engineering
research and practice.

Founded by Dr. Lê Nguyên Khương.

Khương brings structural engineering research, numerical modelling and teaching experience to AIVN360. He is affiliated with the University of Transport Technology and co-leads its Advanced Materials and Intelligent Systems for Infrastructure and High-Speed Railway research group.

Research background at UTT

Affiliation describes the founder’s background and does not imply institutional endorsement of AIVN360.

A focused start. Room to grow.

RailDyn is our first product. Our broader direction includes structural engineering, BIM/IFC, engineering agents and interactive learning labs—future areas shaped by real engineering needs.

04 / Learning

Share the methods
behind the software.

Our existing Vietnamese-language YouTube channel is part of AIVN360’s learning presence. Explore selected discussions on AI systems, evaluation and development practice.

Browse the learning library

AI-assisted audio discussions of credited source material, with links to explore the original work.

Let’s talk engineering

A useful problem
is a good place to start.

For RailDyn, research and education collaboration,
or questions about AIVN360.

lekhuong@aivn360.com