AI-enhanced engineering simulations and software

Mathematics with Purpose

Get in Touch

deepmath offer

Marine engineering and environment

We support marine design and decision-making through CFD, FEM, DES, hydrodynamic modeling, numerical wave tanks, data processing, and AI-enhanced simulations for vessels, offshore assets, and coastal systems.

  • Hydrodynamic and seakeeping analysis
  • Structural integrity and environmental impact studies
  • Design, operation, and logistics optimization

deepmath offer

Offshore renewables

We help teams assess floating wind, floating solar, and offshore energy systems through coupled numerical workflows that connect metocean loads, hydrodynamics, structures, moorings, operations, and performance.

  • Coupled load and dynamic response studies
  • Model calibration, reduction, and digital-twin workflows
  • Design screening and operational optimization

deepmath offer

Aerospace

We build high-fidelity simulation workflows for aerospace problems where physical testing is limited, expensive, or incomplete, combining CFD, FEM, thermal analysis, structural modeling, and numerical interpretation.

  • Aerodynamic, thermal, and structural studies
  • Coupled physics workflow development
  • Design-space exploration and model refinement

deepmath offer

Industrial operations

We model complex operational systems using discrete-event simulation to represent logistics, maintenance, production flows, scheduling, resources, uncertainty, and decision logic over time.

  • Bottleneck detection and throughput analysis
  • Scheduling and resource allocation studies
  • Scenario testing for operational decisions

deepmath offer

Committed to Innovation

We work with companies, startups, and research partners when standard engineering workflows are not enough, combining mathematical modeling, high-fidelity simulation, field-data analysis, AI-enhanced workflows, model reduction, and custom software development.

  • High-fidelity simulation for new engineering concepts
  • Hybrid physics-AI models and reduced-order workflows
  • Custom mathematical tools and engineering software

deepmath offer

FEM

We use finite element methods to model structural behavior, stress, deformation, materials, contact, fatigue, and coupled physical response where engineering confidence depends on numerical depth.

  • Structural integrity and reliability assessment
  • Complex loads, materials, and boundary conditions
  • Coupled physics and simulation workflow automation

deepmath offer

CFD

We develop CFD workflows for complex fluid, thermal, hydrodynamic, and coupled flow problems, turning pressure, velocity, turbulence, and free-surface behavior into practical engineering evidence.

  • High-fidelity fluid and thermal simulation
  • Pressure, velocity, turbulence, and free-surface analysis
  • Fluid-structure interaction and model calibration

deepmath offer

DES

We use discrete-event simulation to model operational systems driven by sequences of events, resources, constraints, scheduling, uncertainty, and decision logic over time.

  • Logistics, maintenance, and production-flow modeling
  • Bottleneck detection and scenario testing
  • Resource allocation and statistics-based digital twins

deepmath offer

AI

We apply AI as a physics-aware layer in engineering simulation, learning from high-fidelity results, accelerating workflows, supporting predictive models, and reducing repetitive engineering effort.

  • Surrogate and reduced-order model development
  • Intelligent meshing and preprocessing support
  • Physics-informed and generative simulation workflows

deepmath offer

Digital twins

We build digital-twin workflows that connect simulation models, operational data, calibration, and AI-driven inference to support monitoring, prediction, and engineering decisions.

  • Simulation and sensor-data integration
  • Monitoring, prediction, and operational decision support
  • Model calibration and faster engineering updates

deepmath technology

deepmesh

deepmesh is our flagship AI-powered meshing technology, designed to automate high-quality mesh generation for CAE and CFD workflows while improving consistency, simulation readiness, and engineering control.

  • AI-assisted mesh generation for CAE and CFD
  • Quality-aware topology and element distribution
  • Workflow integration for simulation readiness

deepmath technology

The Bottleneck

Mesh generation remains one of the most expert-driven and time-consuming steps in CAE and CFD workflows, with mesh quality directly affecting numerical accuracy, solver stability, and computational efficiency.

  • Geometry cleanup, refinement, and boundary-layer decisions
  • Solver stability and numerical reliability
  • Reduced manual correction cycles

deepmath technology

The Shift

deepmesh moves meshing from manual setup toward intelligent generation, combining graph neural networks and transformer-style architectures to learn local geometric detail and global mesh structure.

  • Local geometry and global structure reasoning
  • Engineering-rule-guided automation
  • Self-learning mesh generation workflows

deepmath technology

The Outcome

The goal is to reduce preparation time, improve consistency across studies, support solver efficiency, and give engineers more room to focus on physics, design comparison, and simulation-driven decisions.

  • Shorter mesh preparation cycles
  • More consistent simulation inputs
  • Faster engineering design and validation

deepmath

About Us

We are a deeptech startup founded in 2023, advancing engineering simulations through cutting-edge numerical methods and artificial intelligence. We support innovation across industries including offshore renewables, marine engineering, aerospace, automotive, and energy, delivering high-fidelity or real-time simulations for critical engineering decisions. More here.

At the intersection of advanced simulation and AI-driven engineering software, we develop intelligent tools that go beyond standard workflows. Our flagship innovation is an AI-powered intelligent mesh generator for CAE and CFD. A key enabling technology designed to remove long-standing bottlenecks, accelerate workflows, and improve simulation quality. More here.

With a multidisciplinary team based in France and Brazil and a growing presence in the United States, empowers engineers wordwide to design faster, reduce computational costs, and build more sustainable solutions.

Value Proposition

We help engineering teams build a numerical laboratory around their assets, from simulation-based design to operation. By combining high-fidelity physical models with AI-driven predictive capabilities, enables faster design iteration, asset performance prediction, predictive maintenance planning, and better-informed technical decisions.

The result is a tailored engineering decision-support environment, turning simulations, operational data, and engineering knowledge into actionable insight for each client's specific assets, industries, and constraints.

Physical World diagram

1. Physical World

Real assets, operating conditions, environmental loads, and field constraints define the engineering problem.

Engineering Simulations diagram

2. Engineering Simulations

High-fidelity numerical models translate physical behavior into measurable simulation outputs.

AI-Powered Engineering diagram

3. AI-Powered Engineering

Learning systems connect simulation evidence, design variables, and expert knowledge into reusable models.

Predictive Digital Twin diagram

4. Predictive Digital Twin

A live engineering representation supports faster iteration, operational insight, and robust decisions.

The result is a tailored engineering decision-support environment, turning simulations, operational data, and engineering knowledge into actionable insight for each client's specific assets, industries, and constraints.

Predict outcome diagram

Predict

Future asset behavior through high-fidelity physical and AI-driven models.

Optimize outcome diagram

Optimize

Designs, operating strategies, and maintenance planning with faster engineering iteration.

Decide outcome diagram

Decide

With confidence using tailored insights grounded in simulation, data, and asset constraints.

Founding Team

The founding team, in collaboration with its partners, brings together cutting-edge expertise in mathematical modelling. They saw an opportunity to help the industry in building a better future to society and the environment.

Liad is a naval and oceanic engineer, a researcher for 14 years in computational hydrodynamics, offshore environment and infrastructures. He is involved in the study of fluids using artificial intelligence. He earned his degrees at the École Centrale de Nantes.

Liad Picture

Liad Paskin

Founder & CEO

Naval Engineer

PhD in Hydrodynamics

Bruno is a researcher in computational mechanics for 15 years, with expertise covering a variety of fields, such as structures, fluids and artificial intelligence. Bruno earned his degrees at the École Centrale de Nantes in co-supervision with Lisbon and Barcelona.

Bruno Picture

Bruno Tessaro

Founder & CTO

Mechanical Engineer

PhD in Computational Mechanics

Our Experts

Talent acquisition is a core pillar of our growth strategy. Our multidisciplinary team brings together deep expertise in simulation, machine learning, and software engineering.

Ahmed Sherif

Ahmed Sherif

Lead of Simulations & Business Affairs

PhD in Hydrodynamics

Max Zhuofan

Max Zhuofan

Reinforcement Learning Engineer

PhD in Reinforcement Learning

Eya Oueslati

Eya Oueslati

AI R&D Engineer

Master in Robotics and Vision AI

André Amorin

André Amorin

AI & Data Scientist

Earth and Mathematical Sciences

Ítalo Silva

Ítalo Silva

AI & Data Scientist

Computer Science Student

Felipe Kneip

Felipe Kneip

Simulation Engineer

Mechanical Engineer

Stefan Rosca

Stefan Rosca

Computer Scientist

Expert in Computer Science

Partners

Funding

Logo European Union cofinancing
Logo Region Pays de la Loire
Logo BPI France

Ecosystem

Logo KIVO Incubateur
Logo Station F
Logo Atlanpole
Logo Pole Mer Bretagne Atlantique
Logo Images et Reseaux

Technology

Logo Ansys
Logo Nvidia
Logo Microsoft for Startups
Logo Amazon Web Services

Academic

Logo Centrale Nantes
Logo Coppe UFRJ
Logo Ceris

Development

Logo G and M
Logo Promec

Recruiting

Flag Brazil Icon

BR

Oferta de estágio:

• Simulação de Fluidos e Estruturas.

Download pdf.

Contact

deepmath solutions
SIRET 98216894000012
12 Rue Olivier 44100 Nantes

Website publisher: deepmath solutions, SAS with share capital of 25,000 EUR.

Registered office: 1 RUE DE LA NOE, 44300 NANTES, France.

SIREN: 982 168 940
SIRET: 982 168 940 00012
RCS: 982 168 940 R.C.S. Nantes
VAT number: FR09982168940

Publication director: Liad Paskin, President.

Contact: liadpaskin@deepmath.tech

Hosting provider: GoDaddy, 100 S. Mill Ave, Suite 1600, Tempe, AZ 85281, United States.

Accessibility

deepmath aims to make this website accessible and usable for as many visitors as possible. We are continuing to improve the site's structure, keyboard navigation, responsive behavior, contrast, and media alternatives.

If you experience an accessibility issue or need information in another format, contact us at liadpaskin@deepmath.tech and include the page address, the issue encountered, and the assistive technology or browser you are using where relevant.

Our website does not currently use non-essential cookies, advertising pixels, or cross-site tracking technologies. Hosting, security, or browser-level technologies may process limited technical information that is necessary to deliver and protect the site.

If we introduce analytics, advertising, embedded third-party tools, or other non-essential cookies in the future, we will update this notice and request consent where required before those technologies are used.