Research Scientist - Control & Machine Learning

present of AI; Greater London; Hybrid; GBP 60,000 - 90,000; Full time

Job summary

Gigaton is hiring a Research Scientist to bring learning-based control to real manufacturing plants in London. You will work at the intersection of machine learning and physical systems, modelling plant behaviour from data and deploying methods with engineering teams.

The role emphasizes rigorous experiments, production readiness, and collaboration with operators to ensure safety and performance in live environments.

Qualifications

• Deep research experience in control or machine learning.
• Proficiency in Python and PyTorch for research and deployment.
• Experience applying learning-based control to real industrial plants.
• Ability to design rigorous experiments and baselines, and to communicate results with engineers.

Responsibilities

• Develop and validate learning-based control methods for heavy industrial plants.
• Model plant behaviour from noisy data and optimise under uncertainty.
• Move methods from research into production alongside engineers.
• Collaborate with operators and process engineers to identify valuable problems.

Skills

Control theory

Machine learning

Python

PyTorch

System identification

Education

PhD or equivalent research experience

Job description

Research Scientist - Control & Machine Learning

This is a research role focused on getting learning and control methods to work on real heavy industrial plants. The work covers modelling plant behaviour from noisy data, optimising under uncertainty, and moving methods into production alongside engineers. A strong candidate has deep research experience in control or machine learning, writes solid Python, and cares about how methods perform in a live plant.

How it fits the AI era

The role sits at the point where machine learning meets physical systems, with methods like Bayesian optimisation, model based reinforcement learning and physics informed models guiding real equipment. AI can help find patterns in plant data and suggest better actions, but judgment about physical limits, safety and what a failure means still needs people. Talking with operators and process engineers stays a human skill that shapes which problems are worth solving.

How it builds AI experience

Someone in this role would gain hands on experience applying learning based control to real manufacturing plants, from early research through to deployment and follow up on results. It also means learning how to transfer models between sites and learn while controlling. To build this ahead of time, try open source projects in system identification or Bayesian optimisation using simulated industrial processes.

Worth showing

• Control or system identification research
• Python and PyTorch implementation
• Deployed ML or control algorithms
• Rigorous experiments and baselines

This is our own read, written from the employer's posting and the company's Spotlight. The employer's posting is the source of truth for the role.

About Gigaton

Industrial artificial intelligence platform providing self learning process control to cut fuel costs and carbon emissions in energy intensive manufacturing.