Cloudera and Svitla Systems bring trusted AI to healthcare
Trust is perhaps the most important commodity when it comes to our health. We place enormous trust in clinicians, hospitals, pathology services, pharmacies and the myriad of other services that make up our healthcare system. The ability to protect information, while making it available when it's most needed, is a challenge. With the rising tide of cyber threats – the Office of the Australian Information Commissioner reports that healthcare is one of the most targeted sectors – protecting data is critical.
A significant part of the challenge is that data in healthcare systems is often fragmented. Data is collected in different ways and stored in different systems. That makes creating a single view of a patient's data challenging. But AI can be a boon - if it is used in a secure way that respects data governance and security.
Vini Cardoso, Chief Technology Officer of Cloudera for the ANZ region, highlights this. "This is the biggest challenge we have. One approach is to deliver the information to clinicians in an integrated way, while applying a single governance framework. It's about providing clinicians with the right access to the data, with the consent from the patient."
Trust is built when patients can control who accesses their records. Cloudera's new partnership with Svitla Systems seeks to facilitate that access.
Svitla Systems' Executive Vice President for APAC, Andre Koot, explains, "We are building systems that integrate into the workflow that technicians want to see. The focus is on flexibility and integration, which enables multiple access permissions and renewals. The way we work as a company is to help clinicians execute their workflows and ask them specifically where they need access, and how often that access needs to be recertified, to allow ongoing access to that data."
The goal is for a patient to consult a clinician and receive the most accurate, evidence based diagnosis and treatment plan. This demands a robust ecosystem that empowers physicians with comprehensive, real time patient data while respecting privacy and regulatory constraints. And it requires rigorous data governance and patient centric consent frameworks, so patients can granularly control what information is shared.
Current workflows force patients to repeat lengthy medical histories and undergo duplicate diagnostics when seeking a second opinion, driving unnecessary costs and contributing to clinician burnout. Secure, interoperable data exchange means multiple providers can access a unified patient record - this reduces redundant testing, shortens diagnostic timelines, and improves decision quality.
For senior technology managers, the challenge is to create a scalable, policy driven solution that balance data accessibility with compliance.
Alex Barenboim, Chief Technology Officer of Svitla Systems, says other industries have shown this is possible.
"If you go to a bank, even in a different city or country, you can do the same things because banks have created secure ways to transfer money and share information. There needs to be a way for one hospital to safely share information with another hospital."
While the My Health Record service, operated by the Federal Government, makes some of this possible, there remain challenges inside hospitals and healthcare networks.
"From our point of view, it's important to break down data silos inside a hospital or a practice," says Vini. "By providing a platform that allows clinicians to safely share data across the different applications they use, patient care can be vastly improved and costs can be contained."
Achieving this requires a deep understanding of the data assets being used and the ability to tap into multiple systems. Vini says Cloudera helps organisations to have a full understanding of their data assets by mapping where the data is originated - and its movement - so everybody in the organisation knows where data is coming from.
"This is necessary to ensure the data is trusted," says Vini. "As AI becomes more widely used in healthcare, trusting the data that is used to create and update models is critical. Without trust, AI will not be used, and many potential benefits will be lost."
This is where the approach taken by Svitla Systems comes to the fore. Rather than using LLMs with vast troves of data, the focus is on using smaller models built with specific data.
Alex says, "One reason for this approach is we want to avoid hallucinations as much as we can. We also want to avoid risk of AI getting out of control. This is why we build smaller models to do specific tasks and connect them together only where it makes sense.
We focus on building appropriate controls into the model and ensuring the clinician is in the loop. AI has come a long way in a short time but it's critical to retain humans in the loop."
The future of patient care hinges on a trusted data ecosystem that unites fragmented records and brings robust governance and patient centric consent, while retaining humans at the centre of AI decision making. Cloudera's secure, unified hybrid data platform and Svitla Systems' flexible, workflow aligned access controls enable clinicians to pull a complete, auditable patient picture in real time. This is why they focus on AI built on purpose crafted, smaller models that feed on verified data. This adds predictive insight without hallucinations and ensures clinicians remain the final arbiter.
For senior technology managers in healthcare, it's possible to design scalable, policy driven solutions that balance open, interoperable data flow with strict compliance - so that every doctor can deliver the most accurate, evidence based care when it matters most.