TL;DR - Help us build and scale the systems behind Pixel Systems' media authenticity products, with a focus on deep learning. This role balances research depth with shipping code customers rely on.
What you will do
- Design, train, and improve models for synthetic and manipulated content across images, video, documents, and other media, on large datasets with dedicated hardware compute
- Own parts of our training and inference stack in TensorFlow and PyTorch, from experimentation to monitoring in production
- Partner with product and customers to prioritize what to measure and how to explain results
What we are looking for
- 5+ years shipping ML in production as an engineer, with strong theoretical grounding to explain why a model behaves the way it does
- Owned at least one model end to end: data, training, evaluation, deployment
- Clear communication with both technical and non-technical stakeholders
Nice to have
- Publications, open-source models, or benchmarks you have released
This is Pixel Systems
Pixel Systems builds verification infrastructure for high-stakes decisions. When an image, video, or document arrives from a source a business does not control, Pixel tells the team whether it can be trusted enough to act on: a risk score, the signals behind it, and a record that stands up under scrutiny. Insurers, marketplaces, media, legal, and finance teams use it inside the workflows they already run.
- How we work: a small engineering team in Stockholm where each engineer owns what they build from design through production, with AI-assisted tooling as part of the daily workflow
- What you get: market-rate salary, equity, and work on one of the defining problems of the next decade, keeping evidence trustworthy when anyone can fabricate it
Our tech stack
We build on modern cloud infrastructure and AI-assisted tooling so engineers can move fast, stay in flow, and spend their time on problems worth owning.
- Frontend: React, TypeScript
- Backend: Python, C#, TypeScript
- ML: TensorFlow, PyTorch, ONNX, MLflow, dedicated hardware compute
- Hosting: managed cloud, infrastructure as code
- DevOps & tooling: GitHub Actions, OpenTelemetry, Terraform
How to apply
- Apply in English, our working language day to day and the one you will use most if you join.
- We read every application ourselves and treat each one on its merits. If this role fits, submit through this careers page.
- If the role fits but you do not tick every point, apply anyway. We would rather read your application than miss it.