Founding ML Research Engineer

Generative Group; Greater London; Hybrid; GBP 140,000 - 160,000; Full time

Benefits offered by this job

Equity stake

Job summary

Generative Group is seeking a Founding Machine Learning Research Engineer in London to guide the team’s work from research into robust engineering and scalable model training. This engineer-first role combines ML research with production-ready software practices, operating across PyTorch code, training loops, and performance optimisations.

You’ll train models, build data pipelines, and push systems to scale on large GPUs.

Qualifications

• Direct experience training neural networks at meaningful scale, with strong PyTorch expertise.
• Experience with multi-GPU or multi-node training and GPU or hardware performance optimisation.
• Strong Python and production-quality software engineering practices.
• Typically six or seven-plus years’ relevant experience in a mid-career role.
• High agency, rapid learning ability and comfort working in an ambitious early-stage team.
• A Master’s degree, PhD or specific university background is not essential.

Responsibilities

• Directly train and optimise neural networks, build internal workflows and data pipelines.
• Improve training performance across multiple GPUs, currently up to ~20 GPUs in parallel.
• Work across PyTorch model code, training loops, runtime behaviour, hardware utilisation and experiment workflows.
• Collaborate with AI researchers and domain specialists in a fast-moving early-stage environment and turn ideas into reliable systems.

Skills

PyTorch

Python

Multi-GPU training

Performance optimisation

Production-grade software

Job description

About the company

We are a low-profile, frontier AI research lab developing machine learning systems for extreme physics, including plasma physics and multispecies fluid dynamics. Our technology is designed to accelerate research and development across semiconductors, fusion energy, aerospace, defence and advanced engineering.

We build models from the ground up using high-fidelity simulation and experimental data, enabling faster and more reliable predictions across complex physical systems. Our mission is to create positive-impact AI that helps engineers and researchers explore more designs, parameters and operating conditions than traditional methods allow.

The company is seed-funded and has a small, highly technical team of AI researchers. We value scientific ambition, engineering excellence, rapid learning, direct communication, high agency and practical collaboration. This is an opportunity to help shape the foundations of a new category of AI alongside a high-calibre technical team.

What the job entails

As Founding Machine Learning Research Engineer, you will help take the team’s work from research into robust engineering and scalable model training. This is an engineer-first role with a strong machine learning research component—not a pure infrastructure or platform position.

You will directly train and optimise neural networks, build internal workflows and data pipelines, and improve training performance across multiple GPUs, currently up to approximately 20 GPUs in parallel. You will work across PyTorch model code, training loops, runtime behaviour, hardware utilisation and experiment workflows. Depending on your strengths, you may also contribute to model architecture, attention mechanisms, scientific machine learning and evaluation or reproducibility practices.

You will work closely with AI researchers and domain specialists in a fast-moving early-stage environment, making pragmatic technical decisions and turning promising ideas into reliable systems. The role is predominantly office-based in London, with limited focused home working possible.

Key requirements / What we look for

• Direct experience training neural networks at meaningful scale, with strong PyTorch expertise.
• Experience with multi-GPU or multi-node training and GPU or hardware performance optimisation.
• Strong Python and production-quality software engineering practices.
• [Suggestion: experience with vision modelling, world modelling or comparable research environments.]
• Typically six or seven-plus years’ relevant experience in a mid-career role.
• High agency, rapid learning ability and comfort working in an ambitious early-stage team.
• A Master’s degree, PhD or specific university background is not essential.

Compensation & Benefits

• Base salary: £140,000–£160,000.
• Equity: approximately 0.25%–0.5%, depending on experience and salary.

• London-based, predominantly office-based working, with one or two focused home-working days possible.