Modelling & Machine Learning Lead (Photonics)

About LumiAIres

LumiAIres is at the forefront of photonic AI, developing technology with the potential to reshape how the world processes information. We are a values-led company with a flexible, trust-based culture, offering genuine ownership and impact. Our founding team possesses deep expertise in photonics, neuromorphic computing, and commercial strategy, providing exposure to global investors and strategic partners from day one.

The Opportunity

You will build and own the modelling-and-learning capability at the heart of LumiAIres: the simulation, encoding–decoding, and machine-learning layer that makes light compute. This is a durable company function, creating reusable assets across product generations and customer programs. Your immediate focus will be on our flagship LUMINA program, developing a high-fidelity digital twin and encoding–decoding scheme to create a pre-evaluation environment for integrators.

You will act as the bridge between physics and learning, modelling how light computes within our photonic core and training the models that interpret the results. This role offers significant growth potential, with the possibility to lead LumiAIres' entire modelling and machine-learning function as the company scales.

Responsibilities

The role is structured around four core pillars:

1. Photonic & Optical Simulation

• Model the photonic core and full signal chain using established optical-simulation and numerical wave-propagation methods.
• Build, maintain, and validate the high-fidelity digital twin of the photonic processor.
• Quantify how design and fabrication choices affect computational behaviour, feeding insight back to the photonic design team.

1. Encoding, Decoding & Core Behaviour

• Design and optimize input encoding and output decoding schemes for the photonic platform.
• Characterize how the photonic core behaves as a physical computing medium.
• Co-own encoding–decoding optimization with our academic partners.

1. Machine Learning & Readout

• Develop the readout models and associated training pipelines.
• Build the learning loop that improves inference over time and feeds our software tools.
• Optimize the mapping of machine learning inference onto the photonic hardware.

1. Customer-Ready Twins & Validation

• Harden the digital twin into a pre-evaluation environment for integrators.
• Validate twin predictions against measured silicon in partnership with the Applied Photonics Engineer.

Contract & Progression

• Contract Type: 18-Month Fixed Term Contract
• Location: Glasgow, Scotland (Hybrid)
• Salary: TBC per annum
• Reports To: CTO (currently the CEO)
• Potential for Permanent Role: Yes — Glasgow or Paris (determined within the first 12 months)
• Start Date: To be agreed

Requirements

• An advanced degree (MSc or PhD) in applied mathematics, computational or theoretical physics, electrical engineering, computer science, or a related quantitative field.
• Strong optical and photonic simulation skills, including transfer-matrix and numerical wave-propagation methods (or equivalent).
• Grounding in practical machine learning—model training, evaluation, and pipeline development (Python; NumPy / SciPy and PyTorch / JAX or similar).
• Strong scientific software practice and the ability to turn physics into validated, maintainable code.
• Familiarity with photonic neural networks, optical, or neuromorphic computing is a strong advantage.
• Experience with digital-twin or hardware-in-the-loop modelling, active-learning methods, or working alongside experimental and academic teams is welcome.
• Comfort in an early-stage, fast-moving environment where structure is built, not inherited.

How to Apply

To express interest, please send a CV and a short covering note describing your background in optical or photonic simulation and machine learning, any modelling or digital-twin work you have delivered, and what draws you to photonic computing. Applications and initial conversations are treated in strict confidence.