Skip to content

1000x more efficient physics simulations.

Developing the next generation of accelerated simulations for industrial applications

Sign up for beta access

A next-generation physics solver

We are building a new physics simulation solver written from the ground up integrating novel machine learning research and modern numerical programming approaches. Deployed through a web application and API gateway.


  • Order of magnitude efficiency gains
  • Native deployment to cutting-edge hardware accelerators: GPUs, TPUs
  • Adjoint methods for arbitrary optimisation targets

Expensive simulations bottleneck product development

To reduce reliance on expensive physical testing, engineers lean heavily on physics simulations to optimise their designs. However, many design problems across aerospace, manufacturing and integrated circuits are extremely computationally expensive to simulate, taking from hours to days to run; critically slowing product development.

By developing an order of magnitude more efficient physics solver, we will dramatically increase the productivity of engineers. This will allow engineers to efficiently search for optimal designs and therefore develop better products.


The Team

Our team combines deep expertise across applied math, machine learning and software engineering.

A photo of Laurence Cullen, CEO at Vanellus, standing in front of a hedge.

Laurence Cullen

Previously: Fetch.aiSensityARM

Laurence is a machine learning engineer of 6 years skilled at rapid prototyping and has driven two startups to seed as a first and second employee.

Whilst at Sensity, he led the development of one of the world’s first deepfake detection products from scratch, defining and implementing features in consultation with customers from Reuters and The Guardian. He leads on software design, machine learning and commercialisation.

linkedin_logo   twitter_logo   github_logo   gmail_logo

A photo of Dr Michael Negus, CTO at Vanellus, standing in front of a hedge.

Dr Michael Negus

Previously: University of OxfordDiamond

Michael holds a PhD in Applied Mathematics from the University of Oxford, focused on fusing numerical and analytical methods for modelling droplet impacts.

His expertise lies at the interface between mathematics, physics and scientific computing.

He leads the development of our core accelerated physics solver.

linkedin_logo   twitter_logo   github_logo   gmail_logo   oxford_logo

A photo of Dr Aditi Roy, mathematical engineer and first employee at Vanellus.

Dr Aditi Roy
Mathematical Engineer

Previously: KCL, University of Oxford

Dr Roy joins us from her postdoc at the Oxford Department of Computer Science where she ported cardiac simulation code to run on GPUs.

She did her PhD at King's College London developing a pipeline to identify fibrosis in cardiac walls from MRI images and simulate the electrical activity behind atrial fibrillation.

She brings her knowledge of GPU computing and machine learning to improve our solver capabilities.

linkedin_logo   github_logo   gmail_logo   Google_Scholar_logo


Oscar Holroyd
Computational Mathematician

Currently: University of Warwick

Oscar joins us part-time while finishing his PhD at the University of Warwick on thin film fluid control. He is contributing to developing our core solver with his deep range of programming and mathematical skills.

linkedin_logo      gmail_logo  

About us


Laurence and Michael met at the particle accelerator Diamond Light Source as summer students in 2016, and have been friends since. They moved in together to found Vanellus in the summer of 2022.

They kicked off the business by going through Entrepreneur First, going on to raise a preseed investment round in 2023.

Aditi joined the team in September 2023, and we are hard at work building out our core physics solver, targeting a beta launch later this year.



Stay up to date with our team's progress

Contact us