We are looking for a Machine Learning Engineer

We are looking for a Machine Learning Engineer to help build the intelligence layer of a design platform for hardware: models that understand engineering drawings and CAD together, and generate both. You can take a model from a research paper to a production API, and you care as much about whether it solves the right problem as whether it runs efficiently. This is a high-impact, high-agency role. You'll shape how ML is built and evaluated at Encube, and work directly with customers to make sure what you build matters.

Apply here

Location: Stockholm, Sweden
Employment type: Full-time
Department: Engineering

We are Encube

At Encube, we’re building the agentic platform where hardware teams come together to explore designs, share insights and shape ideas to invent better products faster where 3D is at the center.

Every hardware product is described twice: in the 3D model, and in the 2D drawings that go to the shop floor. We're building ML that works across both. It understands how a drawing relates to its model, and it generates CAD and drawings from what an engineer needs.

What you’ll do

Develop and train models that understand how 2D drawings and 3D models relate to each other.
Build generative models and agents that create CAD models and engineering drawings.
Establish rigorous offline and online evaluation. You'll define what "good" looks like.
Build robust data pipelines and training loops (PyTorch, CUDA).
Ship models into the product, and work with the customers who use them.

Our tech stack

We're pragmatic about technology and pick the best tools for the job. Today that's PyTorch and CUDA, trained on our own GPUs and in the cloud, deployed on AWS. It's early days, and you'll have real impact on the tools we choose.

What we’re looking for

Proven experience taking ML systems from prototype to production
Strong instincts around model performance, deployment and evaluation
Experience with multimodal learning or generative models
A pragmatic, product-oriented approach to engineering
Bonus: graph neural networks, 3D geometry, CAD, or engineering drawings

Why this role

The problem space is exciting. Multimodal, multi-representation and domain-constrained. This is not commodity ML, and many of the problems have no off-the-shelf solution.
You're early. The ML foundation is still being laid. The decisions you make in the next 12 months will shape how Encube works for years.

How we work

Autonomy and trust. High ownership, high agency, minimal bureaucracy. You'll have real agency from day one.
Close to customers. The engineers using Encube send us real drawings and real parts, so you always know whether what you built matters.
Truth-seeking, humble team. We value learning, experimentation and impact over ego. We write things down, propose before we build, and measure before we claim.
High standards. Everyone on the team pulls their own weight.

We look for engineers who think in products, not just systems: people who take ownership, drive clarity when it's missing, and create lift rather than drag.

Sound like you?

Apply below along with your CV, tell us in a few sentences about one system you built: what it did, how you knew it worked, and what you'd do differently.

Apply here

Explore other positions

See all open positions

Frontend Engineer

We’re looking for a Frontend engineer who loves solving complex technical challenges with the end-user in mind. This is a high-impact, high-agency role where you’ll shape our frontend architecture and help the team iterate quickly without compromising quality, reliability or security.

ML Ops Engineer

We are looking for an ML Ops Engineer to own the platform our ML team trains, evaluates and ships models on. Our models learn from engineering drawings and CAD, and they train on GPUs in our office and in the cloud. You'll make that platform fast, reproducible and secure, so the ML team can spend its time on models instead of machines. This is a high-agency role: the foundations are in place, and you'll decide how they grow.