Uploaded September 2026 | Updated November 2026, 5 days ago
Explore the Narnia Labs Engineering AI Leaderboard: leaderboard.narnia.ai
Connect with Mason on LinkedIn: linkedin.com/in/민식-서-12a1aa11b
In this deep dive, Mason Seo, CTO of Narnia Labs, walks through a live demonstration of AslanX Flow, from raw engineering data to a deployed surrogate model for 3D field prediction, built without a machine learning engineer on the team.
Starting from 64 solved CFD runs on vehicle aerodynamics, Mason shows how to build a training workflow from scratch on AslanX Flow, train and compare three model architectures side by side (PointNet, Transolver, and Linear Neural Operator), evaluate performance against held-out test data, and deploy the winning model as a no-code inference application in AslanX Studio. The session ends with a live demo of AslanX's MCP integration, where Claude builds and configures the workflow autonomously from a single natural language prompt.
Mason also covers why the model that ranks first on a public leaderboard is not necessarily the right model for your data, and how to measure that yourself on data at the scale most engineering teams actually have.
If you work in engineering simulation, CAE, or are building surrogate models for manufacturing design, this one is for you.
ONLINE PRESENCE
————————
🌍 My website - jousefmurad.com
💌 My weekly science newsletter - jousef.substack.com
📸 Instagram -instagram.com/jousefmrd
🐦 Twitter - https://x.com/Jousefm2
CONTACT:
————————
If you need help or have any questions or want to collaborate feel free to reach out to me via email: support@jousefmurad.com
#EngineeringAI #SurrogateModeling #ManufacturingAI #CFD #AslanX
Disclaimer: Some of these links are affiliate links that make me earn a small commission when you make a purchase at no additional cost.
Explore the Narnia Labs Engineering AI Leaderboard: leaderboard.narnia.ai
Connect with Mason on LinkedIn: linkedin.com/in/민식-서-12a1aa11b
In this deep dive, Mason Seo, CTO of Narnia Labs, walks through a live demonstration of AslanX Flow, from raw engineering data to a deployed surrogate model for 3D field prediction, built without a machine learning engineer on the team.
Starting from 64 solved CFD runs on vehicle aerodynamics, Mason shows how to build a training workflow from scratch on AslanX Flow, train and compare three model architectures side by side (PointNet, Transolver, and Linear Neural Operator), evaluate performance against held-out test data, and deploy the winning model as a no-code inference application in AslanX Studio. The session ends with a live demo of AslanX's MCP integration, where Claude builds and configures the workflow autonomously from a single natural language prompt.
Mason also covers why the model that ranks first on a public leaderboard is not necessarily the right model for your data, and how to measure that yourself on data at the scale most engineering teams actually have.
If you work in engineering simulation, CAE, or are building surrogate models for manufacturing design, this one is for you.
ONLINE PRESENCE
————————
🌍 My website - jousefmurad.com
💌 My weekly science newsletter - jousef.substack.com
📸 Instagram -instagram.com/jousefmrd
🐦 Twitter - https://x.com/Jousefm2
CONTACT:
————————
If you need help or have any questions or want to collaborate feel free to reach out to me via email: support@jousefmurad.com
#EngineeringAI #SurrogateModeling #ManufacturingAI #CFD #AslanX
Disclaimer: Some of these links are affiliate links that make me earn a small commission when you make a purchase at no additional cost.










