Careers
The virtual expert in the clinic
We are building kaiko.w – “double you” – an AI colleague for clinicians. Giving clinicians greater capacity for care.
Healthcare isn't a fixed sequence of steps.
A new lab result, scan or note can change the next step entirely, and clinicians adapt constantly as they move across records, images, and conversations. Most AI tools only tackle fragments of this reality, handling fixed tasks in isolation.
Yet the real promise of AI is broader – systems that can reason, adapt, and extend expertise – and that is the path we are taking. kaiko.ai is building an agentic assistant that can mirror how clinicians actually work, supporting both routine administration and expert decision–making. To make this possible, kaiko.w learns not only from static data, but also inside a virtual clinic designed to expose AI to the same kinds of workflows doctors face every day.
Already in hospitals
kaiko.w is designed for workflows, not one–off tasks. Integrated with EHRs and other hospital software, it helps with everyday work: structuring referral letters, preparing consultations, and grounding searches in published guidelines. It also understands longitudinal data, following patients across time and connecting records, images, and results. In oncology, this means supporting tumor tracking and related advanced workflows. With custom templates, the assistant adapts to each department and is evaluated directly with clinicians to ensure safe, reliable performance.
Already in hospitals
kaiko.w is designed for workflows, not one–off tasks. Integrated with EHRs and other hospital software, it helps with everyday work: structuring referral letters, preparing consultations, and grounding searches in published guidelines. It also understands longitudinal data, following patients across time and connecting records, images, and results. In oncology, this means supporting tumor tracking and related advanced workflows. With custom templates, the assistant adapts to each department and is evaluated directly with clinicians to ensure safe, reliable performance.
Built in the AI Health Lab
The AI Health Lab is the engine behind kaiko.w. It is a virtual clinic, a reinforcement learning environment where the assistant safely learns workflows, practices tasks, and refines its reasoning before deployment. Here, kaiko.w develops decision–making, tool–use, and adaptation skills in a controlled, testable setting. Training progresses in stages: starting in simulation, moving into shadow mode, and ultimately working under clinician supervision. Each stage generates feedback tied to daily practice, equipping kaiko.w with the reasoning it needs to operate reliably in the hospital. At the core are agentic reasoning models trained for oncology, integrating text, imaging, pathology, and genomics. Because kaiko.w is tool–agnostic, it can also orchestrate validated third–party models and hospital software through a single trusted interface. By “opening” this virtual clinic we give AI a safe, testable path to extend human expertise.
Extending clinical expertise
Each deployment strengthens kaiko.w and moves us closer to our vision: a “double you” for every clinical team – giving clinicians back more time, reducing burnout, and extending expert-level reasoning across care. By combining frontier research with continuous validation in hospitals, we are laying the foundations for the operating layer through which AI will enter medicine. Practical today, transformative over time.
The AI Health Lab is the engine behind kaiko.w. It is a virtual clinic, a reinforcement learning environment where the assistant safely learns workflows, practices tasks, and refines its reasoning before deployment. Here, kaiko.w develops decision–making, tool–use, and adaptation skills in a controlled, testable setting. Training progresses in stages: starting in simulation, moving into shadow mode, and ultimately working under clinician supervision. Each stage generates feedback tied to daily practice, equipping kaiko.w with the reasoning it needs to operate reliably in the hospital. At the core are agentic reasoning models trained for oncology, integrating text, imaging, pathology, and genomics. Because kaiko.w is tool–agnostic, it can also orchestrate validated third–party models and hospital software through a single trusted interface. By “opening” this virtual clinic we give AI a safe, testable path to extend human expertise.
Extending clinical expertise
Each deployment strengthens kaiko.w and moves us closer to our vision: a “double you” for every clinical team – giving clinicians back more time, reducing burnout, and extending expert-level reasoning across care. By combining frontier research with continuous validation in hospitals, we are laying the foundations for the operating layer through which AI will enter medicine. Practical today, transformative over time.
We are developing kaiko.w hand in hand with leading healthcare institutions, ensuring every step is clinically validated and grounded in real–world practice.
Since 2021, kaiko.ai has partnered with the Netherlands Cancer Institute (NKI–AVL), one of Europe’s leading cancer centers. Together we have deployed kaiko.w in live clinical workflows, where clinicians are already using it to support their daily work. Building on this foundation, we are now expanding to hospitals across Europe, ensuring the assistant evolves alongside clinicians and their real–world needs.
We are building the operating system for AI in healthcare
Join us in shaping how AI becomes part of everyday clinical practice.
View careers