Senior/Staff RL Engineer - ML R&D

Das ist der Job

You diagnose these at root cause, fix them, and contribute back upstream where you can.

Darum lohnt es sich

On top of this foundation, we are developing specialized diagnostic agents in areas such as oncology, radiology, and pathology. /p pWe build in close collaboration with leading hospitals and research centers, including the Netherlands Cancer Institute (NKI). kaiko is a well-funded company with a growing international team, operating from Zurich and Amsterdam. /p h3About The Role /h3 pKaiko trains its own foundation models for clinical work on a dedicated GPU cluster.

We've built a team of international experts where your work has a direct impact.

Here's what we value: /p ul liOwnership: You'll have the autonomy to set your own goals, make critical decisions, and see the direct impact of your work. /li liCollaboration: You'll approach disagreement with curiosity, build on common ground, and create solutions together. /li liAmbition: You'll be surrounded by people who set high standards for themselves and others, who see obstacles as opportunities, and who are relentless in their work to create better outcomes for patients. /li /ul pIn addition, we offer: /p ul liAn attractive and competitive salary, a good pension plan, and 25 vacation days per year. /li liGreat offsites and team events to strengthen the team and celebrate successes together. /li liA EUR 1000 learning and development budget to help you grow. /li liAutonomy to do your work the way that works best for you, whether you have a kid or prefer early mornings. /li liAn annual commuting subsidy. /li /ul pOur interview process is designed to assess mutual fit across skills, motivation, and values.

It typically includes the following steps: /p ul liScreening call: A short conversation to align on your motivation, professional goals, and initial fit for the role. /li liTechnical interview with offline assignment: A deep dive into your problem-solving approach through a technical challenge, case study, or role-specific scenario. /li liOnsite meeting (optional): You'll meet team members across functions to explore collaboration dynamics, team fit, and day-to-day context. /li liFinal executive conversation: A discussion with a member of the executive team focused on long-term alignment and shared expectations for impact. /li /ul /p #J-18808-Ljbffr ppKaiko is building a next-generation agentic clinical AI assistant that helps clinicians reason across patient data, guidelines, and diagnostics. /p pHealthcare decisions are rarely made by a single person or from a single data source. kaiko's assistant maintains longitudinal patient context across encounters, clinicians, and institutions, enabling collaboration, second opinions, and complex diagnostic workflows.

The system is designed to operate safely in real clinical environments, with human oversight, auditability, and regulatory alignment at its core. /p pOur assistant core supports broadly applicable clinical tasks such as patient data navigation, guideline interaction, multimodal interaction (chat and voice), and care coordination.

RL is the engine driving alignment, reasoning, and agentic capability across our stack. /p pYou own the RL training infrastructure end-to-end: the distributed training stack, the reward pipelines, and the experiment infrastructure that lets researchers iterate fast.

The hard problems are real, reward hacking and objective-level instability, entropy collapse as policies converge prematurely, sparse and delayed rewards that make credit assignment across long reasoning traces extremely difficult, and exploration failures on hard problems where the model rarely samples a correct trace and learning stalls entirely.

You also explore new algorithms - from policy gradient variants and offline RL to agentic RL with tool use - and bring what matters into production. /p pYou will be based in either The Netherlands or Switzerland, with the expectation of spending at least 50% of your time at the office. /p h3Some areas of responsibility /h3 ul liOwn the RL training stack end-to-end and keep it scaling cleanly across large MoE models and long contexts. /li liBuild and maintain reward pipelines: verifiable reward signals, LLM-based reward models, and reward shaping strategies for complex clinical reasoning tasks. /li liDebug training instabilities at root cause — reward hacking, entropy collapse, credit assignment failures, gradient issues — and ship fixes, not workarounds. /li liExplore new RL algorithms and reward designs; run controlled experiments and translate promising results into the main training stack. /li liScale runs across more nodes, longer contexts, and more complex parallelism as models and tasks grow. /li liContribute upstream to open-source frameworks when you find bugs or missing features. /li /ul h3About You /h3 ul liDeep hands-on experience with RL training systems: you have shipped and scaled RL or post-training runs, not just run tutorials. /li liFluent in at least one distributed training framework at a level where you can read the source and debug silent failures. /li liStrong understanding of core RL challenges: reward hacking, credit assignment, exploration, entropy collapse, sample efficiency — and practical ways to address them. /li liComfortable at the intersection of research and engineering: you read papers, implement ideas, and know when something is worth productionising. /li liExcellent software engineering: clean Python, typed code, reproducible experiments, good test coverage. /li liIndependent operator: you don't need prescribed task lists; you take a system from "running" to "stable, fast, and understood." /li /ul h3Nice to have /h3 ul liExperience with verifiable reward signals or LLM-as-judge reward pipelines. /li liFamiliarity with inference serving systems as part of an RL rollout loop. /li liExperience with MoE training and the additional complexity it introduces. /li liContributions to open-source training frameworks. /li liExposure to agentic or tool-use RL — web search, code execution, multi-step reasoning. /li liHealthcare or regulated-deployment context. /li /ul pAt kaiko, we believe the best ideas come from collaboration, ownership, and ambition.