Deeplabs provides Deepyt®, a no‑code AI platform that converts historical CAD and CAE data into compact latent parameters using autoencoders, enabling instant performance prediction and automated design optimization. The tool integrates with existing CAD workflows to eliminate manual meshing and reduce the need for repeated CFD/FEM simulations, allowing engineers to generate and evaluate design alternatives in real time. It targets mechanical and aerospace engineers, CAE analysts, and product development teams seeking faster, data‑driven iteration.
Funding
Funding not disclosed
Founders
Product
Problem
Industrial engineering design cycles often rely on manual CAD modeling, time‑consuming meshing, and repeated CFD/FEM simulations, which delay product development and limit the ability to explore alternative designs. Additionally, large volumes of existing CAE data are underutilized, preventing data‑driven performance improvements.
Solution
Deeplabs offers Deepyt®, a no‑code AI platform that ingests historical CAD and CAE data to create compact latent representations of 3D designs using autoencoders. These latent parameters enable rapid performance prediction and automated optimization without the need for manual meshing or extensive simulation runs. Engineers can generate design alternatives, evaluate them instantly, and iterate toward optimal solutions, significantly shortening development timelines. The platform integrates with existing CAD workflows and supports both engineering‑focused and data‑science‑focused use cases, allowing users to build custom machine‑learning models without programming. By leveraging AI, Deepyt® transforms raw design data into actionable insights, improving product performance and reducing engineering effort.
Target Audience
Primary customers are mechanical and aerospace engineers, CAE analysts, and product development teams in manufacturing firms that require faster design iteration and data‑driven optimization.
Features
- Autoencoder‑based compression of 3D CAD models into low‑dimensional latent parameters
- Instant performance prediction for CFD/FEM outcomes using trained ML models
- Automated design optimization loop that proposes geometry variations and evaluates them in real time
- No‑code interface for building, training, and deploying machine‑learning models on engineering data
- Seamless integration with standard CAD tools and CAE pipelines for clean‑up, meshing, and simulation
- Support for both engineering (turbomachinery, CAE analysis) and data‑science projects within the same platform