Machine Learning Engineer - Agentic AI & Reinforcement Learning

Viridien () is an advanced technology, digital and Earth data company that pushes the boundaries of science for a more prosperous and sustainable future. With our ingenuity, drive and deep curiosity we discover new insights, innovations, and solutions that efficiently and responsibly resolve complex natural resource, digital, energy transition and infrastructure challenges. Are you interested in building AI systems for real-world scientific problems?

Viridien is expanding its capabilities in agentic AI, large language models and reinforcement learning for complex scientific and enterprise workflows. We operate global high-performance scientific computing infrastructure delivering more than 700 petaflops and executing over one million scientific computing jobs each day across areas including geophysical imaging, materials science, biotechnology and AI-driven research. We are looking for a hands-on Machine Learning Engineer to help build and evaluate intelligent agent systems, turning emerging AI methods into robust and adaptive solutions for scientific applications.

Who you'll work with

As a member of our AI Lab, you'll collaborate with senior machine learning engineers, software developers, HPC specialists and scientific domain experts across Viridien. Learn more about the team on our website, in a recent blog post featuring one of our senior machine learning engineers and from our Voices of Viridien.

What you'll work on Designing and developing single-agent and multi-agent systems.

Implementing and evaluating capabilities such as reasoning, planning, task decomposition, memory, tool use and agent collaboration.

Integrating LLMs, vision-language models, retrieval systems and domain-specific tools.

Experimenting with fine-tuning and adapting foundation models using supervised learning, preference optimisation and reinforcement learning where appropriate.

Developing training environments, reward functions, graders and feedback mechanisms.

Designing evaluations, guardrails and monitoring approaches for agent reliability.

Helping turn successful prototypes into APIs, services and reusable software components.

What we look for Essential

Degree in computer science, artificial intelligence, machine learning, engineering or a related discipline, or equivalent practical experience.

Good Python programming skills and an understanding of software engineering practices.

Experience building machine-learning, generative-AI or LLM applications, gained through employment, industry or research internships, applied research or significant academic projects.

Experience with PyTorch, JAX, Tensor Flow or a similar framework.

Familiarity with core agentic AI concepts such as tool calling, planning, memory and workflow orchestration.

An understanding of reinforcement learning, preference optimisation or sequential decision-making.

Some experience evaluating models or implementing ideas from machine-learning research.

Desirable Familiarity with multi-agent systems using Lang Graph, Auto Gen, smolagents or similar frameworks.

Some exposure to LLM fine-tuning or preference and reinforcement-learning approaches, such as DPO, GRPO, RLHF or RLAIF.Experience designing reward functions, graders or agent training environments.

Familiarity with Model Context Protocol or other approaches to tool integration.

Exposure to training or deploying open-source language or vision-language models using GPUs, including distributed training.

Experience applying AI to scientific, engineering, energy or geoscience workflows.​Why work with us Competitive salary commensurate with experience

Highly attractive bonus scheme

Initial 22 days annual leave with future increases, complemented by a flexible buying and selling holiday programme

Hybrid working:…