Job Description

About kaiko

kaiko.ai is a clinical AI research lab in Zurich and Amsterdam. We train our own large-scale frontier-class models end to end, on our own compute. The full stack is in scope, from training infrastructure and post-training algorithms to data generation, evaluation, and production inference.

Our models are developed and deployed at leading European hospitals, and that deployment feeds back into the next training cycle.

We're hiring interns across the lab. You'll work on the real problems of training and serving these models in a regulated clinical setting.

The Opportunity

As an ML Research Intern, you will be embedded in our core research team for six months, working alongside PhD researchers and senior engineers on problems that directly shape our clinical foundation model. You will own real work, contribute ideas, and see your output reflected in our system.

Depending on your background and preferences, your project may sit in one of the following areas:

• Data: Synthetic clinical data generation, multimodal data pipelines, curation and quality filtering.

• Evaluation: Clinical benchmarks, agentic eval design, measurement infrastructure for the production stack.

• Scaling: Continued pre-training on clinical corpora, post-training algorithms, distributed training at frontier scale.

• RL: Agentic RL on real clinical workflows, reward modeling, RL infrastructure.

• Acceleration: Inference & Training optimisation, serving stack, quantisation and throughput tuning.

You will discuss and agree on your specific focus during the interview process, based on mutual fit and current priorities.

Who You Are

We are looking for MSc students or recent graduates in machine learning, computer science, or a related quantitative field. We care more about sharp thinking, high agency, and genuine curiosity than a perfect CV.

You should bring:

• Strong programming in Python and comfort with ML frameworks (PyTorch or equivalent).

• Hands-on experience with ML research or engineering, ideally in training, post-training, RL, data, or evaluation. Coursework, thesis, internship, or open-source all count.

• Clear technical communication, in design docs and in discussion with researchers, engineers, and clinical scientists.

• Motivation to work on hard problems with real ownership. Intern projects ship into the live training and serving stack.

The specific area depends on the project. None of these are required upfront.

Why kaiko

• Work at the frontier: you will be in the room with PhD researchers and engineers who have built production ML systems at top labs and institutions. Expect to be challenged and to grow fast.

• Ownership: you will set goals, make decisions, and present your findings directly to the research team, not to an intern coordinator.

• Conversion pathway: strong interns are actively considered for full-time roles. This is how we build the team.

• Fair compensation an internship salary of CHF4000/EUR3000 gross per month

Interview Process

We keep it fast. The process typically involves three steps:

• Screening call (30 min): motivation, background, and initial fit.

• Technical assessment: a short, focused problem relevant to your area, reviewed together in a follow-up conversation.

• Onsite or video interview: meet the team, go deeper on your thinking, and align on project fit.

We review applications on a rolling basis. If you are finishing your degree this summer and want to start in September, apply now.