Job Type: Permanent Full Time, On-site
The Company
Luffy AI is an exciting high-tech startup developing adaptive neural networks for industrial control & robotics. Luffy specialises in “Micro AI” controllers that can be trained in simulation on digital twins and successfully transferred into real world systems. The Micro AI controllers can be deployed on existing controller hardware with a small footprint and no internet connection.
Our networks use neural plasticity to learn the dynamics of the equipment they are placed in and continue to adapt long after training. These innovations allow us to overcome the difficulty of applying AI in control system applications.
Our transformational AI technology allows process industries and manufacturers to improve productivity and save energy, and allows industrial automation vendors to simplify and extend the operating envelope of their machines. This revolutionary control technology is a key enabler of Industry 4.0, with huge potential in foundation industries such as metals, glass and cement manufacturing, as well as in automation sectors such as electric motors and robotic systems.
The Role
We are looking for a skilled and experienced Simulation & Modelling Engineer to join our multidisciplinary delivery team. The team is responsible for building our suite of AI products for control and optimisation of industrial processes and equipment. We are looking for someone with deep practical experience in reduced order modelling of physical, chemical or industrial systems and processes who is looking to help us solve real world problems with artificial intelligence.
You will start work immediately on a grant funded research project into AI control of unmanned aerial vehicles (UAVs), working in partnership with a University. You will also work on commercial contracts that Luffy AI is pursuing in various areas of industrial controls (AI control of industrial motors, furnaces, chemical processes, etc…).
When training AI (neural network) controllers for these applications, we require digital twin models that expose the AI controllers to a representative approximation of the system dynamics and phenomena encountered.
AI controllers are trained on these digital twins using Reinforcement Learning techniques, and in some instances, the digital twin model is also a standalone product offering. These digital twin models are built on in-house engines, which are configured for the application under development.
This dynamic role will require you to have a strong interest in modelling real world systems and software engineering, with broader interests such as AI, high performance computing, and mechanical or electrical systems. You will be able to work with a high degree of autonomy, with the ability to engage with the mathematics and physics literature, seeking out domain expertise as required. You care deeply about software standards and enjoy working on challenging research problems, sometimes requiring experimentation.
Key Responsibilities:
* Implement mathematical models of real-world systems.
* Develop our core engines and reference models with a focus on computational performance.
* Conduct data analysis, model calibration, verification and validation.
* Assist the AI team with digital twin development. Design/develop simulation engines taking into consideration numerical stability, performance, and accuracy.
* Capture relevant physics and chemistry underpinning a customer’s application, developing requirements for the model engines.
* Engage with customers and partners to gather data and feedback need to develop, improve and calibrate models
* Maintain good code quality standards across our model code bases.
* Develop visualisation tools for digital twins and physics engines, alongside other developers.
Required Qualifications and Experience:
* University degree in a relevant area of engineering/science (physics, chemical engineering, robotics, maths, computer science) or an equivalent qualification. An advanced degree may be an advantage but is not required.
* 2+ years of relevant experience in one of digital twin development, and physical, mechanical or chemical process modelling, or other physical simulations balancing performance and fidelity.
* Good understanding of the numerical techniques related to physical codes, e.g. differential equation solvers, numerical integration schemes, statistical sampling, finite element methods.
* Strong mathematical skills, including vector calculus, linear algebra, 3D geometry.
* Experience of developing foundational models: system analysis, mathematical model algorithm development, parametric identification.
* Strong programming skills in Python, experience with C++ or Rust desirable.
* Ability to produce visualisation tools that support simulation models (graphing and 3D visualisations, Python/JS).
* Must be able to autonomously research and learn about new systems/processes. Be able to make and justify approximations relevant to the simulation fidelity required.
* Good understanding of code profiling, performance and numerical optimisation techniques.
* Familiarity with testing frameworks and continuous integration.
* A desire to help other people solve their problems.
* Good communication skills.
Behaviours and personal characteristics:
* Passionate about physics / robotics / chemical engineering and about writing robust, efficient code that is well thought out and tested.
* Fast learner, comfortable picking up new technologies and techniques.
* Product focused, good prioritisation skills.
* T-shaped person, deep expertise in a few areas but able to be a generalist when needed.
* Team player.
* Strong organisational skills.
Benefits:
* Full time salary of £45k - £65k, depending on experience and capability assessment during the interview process.
* EMI share options scheme.
* 25 days annual leave, plus bank holidays