Simulation Platform & ML Engineer

Turn cutting-edge simulation and ML research into software that engineers can actually use.

At GREEN14, we are building a new generation of simulation technology for complex industrial processes.

Our simulation team combines multiphysics models, machine learning and experimental process data to understand systems such as plasma reactors and high-temperature metallurgical processes.

We can already build sophisticated models and fast surrogates. Now we want to turn that capability into robust software that can eventually be used beyond our own research environment.

That requires a different kind of engineering.

What we are building

GREEN14 started five years ago with a mission to help Europe develop emission-free production of critical raw materials using hydrogen plasma.

That journey gave us something increasingly valuable: our own experimental infrastructure.

Today, we operate a full-scale plasma reactor in Stockholm. Alongside it, we have built capabilities in multiphysics simulation, internal solvers, AI and surrogate modelling.

Our ambition is to connect these worlds:

Experiment physics simulation surrogate optimization experiment

But a great model sitting in a research notebook isn't a product.

Models need data. Data needs pipelines. Experiments need to be reproducible. Software needs APIs. Models need to be versioned and tested. Computation needs to be fast and reliable.

That is where you come in.

Your role

You will build the software and computational infrastructure connecting our physics, AI and experimental data.

You will work directly with the scientists and engineers developing our models and take research code toward software that other engineers can reliably use.

A typical problem might start with JAX code developed by one of our researchers. Your job is to ask:

How do we make this reliable, testable, reproducible, fast and usable by someone who didn't write it?

You will have substantial ownership over the technical foundations of our simulation platform.

What you will do

You will design and implement robust backend architectures connecting sensor data, physics models, ML models and simulation workflows.

You will turn JAX-based simulation and ML research code into efficient, maintainable and production-ready software.

You will build pipelines connecting experimental and sensor data model evaluation fast simulation and surrogates visualization and optimization.

You will develop infrastructure for model training, checkpointing, evaluation, experiment tracking and reproducibility.

You will structure repositories, APIs, packages and development workflows that can support a growing engineering organization rather than a single researcher.

You will work with databases and data platforms capable of handling large volumes of simulation, experimental and model data.

You will help establish testing, CI/CD, versioning and monitoring practices appropriate for scientific software.

You will work with our simulation and AI engineers to identify computational bottlenecks and improve performance.

As our products mature, you will contribute to deployment, cloud infrastructure and the architecture required to expose simulation capabilities to industrial users.

And you will work closely with product and frontend engineering to turn highly complex computational capabilities into software that feels surprisingly simple to use.

Who you are

You have experience taking technically complex software from prototype to something other people can reliably use.

You might come from software engineering, ML engineering, scientific computing, data engineering or a similar background.

We are particularly interested in experience with:

• Python and modern software engineering practices
• JAX or other numerical/ML frameworks
• Production-grade backend architectures and APIs
• Data pipelines, databases and modern data platforms
• ML infrastructure and MLOps
• Experiment tracking, model and version management and reproducible workflows
• Testing, packaging, CI/CD and repository architecture
• Performance optimization and computationally intensive workloads

Experience with Docker, cloud infrastructure, Kubernetes, GPU computing, FastAPI, MLflow, Weights & Biases, DVC or similar technologies is useful.

Experience deploying or optimizing JAX workloads, scientific computing or simulation software is particularly relevant.

But we don't expect you to tick every box.

The kind of person we are looking for

You don't need to be a plasma physicist.

What matters is that you enjoy working at the boundary between scientific computing, machine learning and real-world software engineering.

You like understanding what researchers are trying to achieve, but your instinct is to turn clever prototypes into systems that are robust enough for other people to depend on.

You care about architecture, but you also write code.

You care about performance, but you also care about maintainability.

And when five different research scripts solve versions of the same problem, your natural reaction is to figure out what the underlying platform should look like.

You should also be curious about the physical world behind the software.

Our code ultimately represents real plasma, real particles, real materials and real experiments. You don't need to understand all of the physics on day one, but you should want to.

Why GREEN14

Many software engineers working in simulation never see the physical system their code represents.

Here, the reactor is downstairs.

Our process engineers can run an experiment. Sensors generate data. Our simulation team can compare the result with the model. The model can improve. And you can build the infrastructure that makes that loop faster, more reliable and eventually scalable to other processes.

If we succeed, the software you build won't just support one reactor.

Our ambition is to create a platform that can help engineers understand, optimize and scale complex high-temperature processes across industries.

How we work

You'll join a team of serious engineers and scientists who care deeply about getting things right.

We're international, curious and direct. We challenge each other's thinking, share what we know and change our minds when the data tells us to.

There is very little hierarchy and no room for corporate politics. Good ideas matter more than titles, and we expect people to speak up when they think we're wrong.

We're ambitious, but low on ego. We help each other get better because the problems we're trying to solve are genuinely hard — and none of us can solve them alone.

We work hard, but flexibly, and we value having a life outside work.

The practical stuff

Compensation consists of salary and pension together with qualified employee stock options.

We are based at Sweden's leading technical university, a few minutes from central Stockholm. There is a reason for that location: we want the people building the software close to the reactor, the experiments and the scientists using it.

If you like the idea of turning physics, AI and experimental data into software that could change how industrial processes are developed and optimized, we'd like to hear from you.

The position will remain open until we find the right person.