Revolutionizing Radiology: A Conversation with HOPPR.ai's Dr. Khan Siddiqui

In this exclusive interview, Dario Priolo sits down with Dr. Khan Siddiqui, CEO and founder of HOPPR.ai, to discuss how AI foundation models are transforming medical imaging and patient care. Click here to watch the full recording.

Dario Priolo: Thank you for joining me today, Dr. Siddiqui. HOPPR has been making significant waves in healthcare AI. Could you share your vision for what HOPPR is trying to accomplish?

Dr. Khan Siddiqui: Thanks for having me, Dario. At HOPPR, we're focused on a fundamental challenge in healthcare—how to accelerate AI development in medical imaging while keeping patients at the center of everything we do. Our origin story actually dates back to a memo I wrote at Microsoft in 2012, thinking about how we could build tools to help radiologists amid growing burnout and volume overload.

The core problem we're solving is simple but critical: traditional AI development in medical imaging takes 18 to 24 months from concept to deployment. We've created a platform that compresses that timeline to just three months, while ensuring everything remains secure, compliant, and focused on creating patient value.

Dario Priolo: That's impressive. Could you walk us through how this actually impacts patient care?

Dr. Khan Siddiqui: Absolutely. Consider something as routine as a mammogram. Today, about 15% of mammograms need to be repeated, often because of technical issues like motion artifacts or incomplete coverage. When that happens, a patient typically leaves the facility, and days or weeks later receives a letter saying they need additional imaging—with no explanation.

Imagine the anxiety that creates. But with our technology integrated into the imaging center's workflow, those issues can be identified immediately. A patient can get additional imaging while still at the facility, eliminating both the anxiety and the inefficiency of scheduling a return visit.

Dario Priolo: So it's about catching problems in real-time rather than after the fact?

Dr. Khan Siddiqui: Exactly. And the applications extend far beyond mammography. In emergency settings, our platform can help prioritize critical cases. Imagine a busy trauma center receiving multiple patients simultaneously—the AI can flag the most urgent findings, like an intracranial hemorrhage, within seconds of images being acquired. Those saved minutes can make profound differences in patient outcomes.

Dario Priolo: Most AI companies in healthcare sell directly to hospitals or imaging centers. I understand HOPPR takes a different approach?

Dr. Khan Siddiqui: We do. We've chosen a B2B model where we provide our foundation models through APIs to developers, PACS vendors, and other AI companies. We're not trying to replace existing systems—healthcare has enough disruption. Instead, we're enabling those systems to become AI-powered.

Think of us as "Intel Inside" for medical imaging AI. Our partners bring specialized expertise in specific clinical domains, and we provide the secure, compliant foundation that accelerates their development.

Dario Priolo: That's an interesting model. What led you to take that approach rather than developing applications directly?

Dr. Khan Siddiqui: It comes down to understanding how healthcare technology actually gets adopted. Most facilities already have viewing applications and PACS systems they've invested millions in. If you try to replace those systems, you face enormous resistance—it's a replacement strategy rather than an acquisition strategy.

By providing our technology through APIs, we allow existing systems to be enhanced rather than replaced. A radiologist doesn't need to learn a new interface or open another application—the AI capabilities are seamlessly integrated into tools they already use.

Dario Priolo: Security and compliance are major concerns in healthcare. How does HOPPR address those challenges?

Dr. Khan Siddiqui: We address four key risks. First, contractual risks—ensuring our agreements with data partners are comprehensive and appropriate. Second, technical risks—making sure data is secure, doesn't leave the platform, and complies with HIPAA and other regulations. Third, operational controls—strictly limiting who has access to environments. And fourth, reputational controls—HOPPR doesn't sell data, that's not our business model. Our model is compute.

This comprehensive approach allows our partners to focus on their clinical expertise rather than spending 6-9 months building compliance and security infrastructure. It's a win for them and ultimately for patients.

Dario Priolo: You've mentioned the idea of foundation models several times. Could you explain what that means in the context of medical imaging?

Dr. Khan Siddiqui: Foundation models are essentially large AI models trained on vast and diverse datasets that can then be fine-tuned for specific tasks. Most people are familiar with ChatGPT, which is a foundation model for text. Our model, which we call "Grace," is similar but specializes in medical imaging.

What makes our approach unique is that we preserve the full diagnostic integrity of images. While many AI tools downscale to just 256 shades of gray, our proprietary vision transformers maintain up to 65,000 shades—ensuring no loss of valuable diagnostic information. This enables the detection of subtle abnormalities that might otherwise be missed.

Dario Priolo: How do healthcare organizations typically engage with HOPPR? Walk me through that process.

Dr. Khan Siddiqui: Most organizations work with us through our partners. For example, a PACS vendor might integrate our foundation models into their existing system, allowing their customers to access advanced AI capabilities without changing their workflow.

Let me give you a concrete example. We're working with a large PACS manager that's deploying multiple models for negative triage of chest X-rays. About 80% of outpatient chest X-rays are normal, so if you can identify and prioritize those versus the positive ones, you gain tremendous efficiency. This is especially valuable for on-call scenarios where you're paying premium rates to teleradiology companies.

The PACS company comes to us with the problem they're trying to solve, we work together to fine-tune our foundation models for that specific use case, and then they implement it through an API integration. The end result is that their product is now AI-enabled, delivering more value to their customers.

Dario Priolo: That's fascinating. Can you share some insights about the fundraising journey for HOPPR? I imagine operating at the intersection of deep tech and healthcare presents unique challenges.

Dr. Khan Siddiqui: Fundraising is always hard, especially when you're in a space that bridges multiple domains. The challenge is finding partners who understand both the technical requirements of building large foundation models and the complexities of healthcare—regulatory pathways, reimbursement models, and clinical workflows.

Typical healthcare investors often don't appreciate the investment needed for deep tech, while deep tech investors may not fully grasp the regulatory aspects of healthcare. The key is doing the homework to find partners aligned with your vision and philosophy.

We've been fortunate to find strategic investors who believe in our mission and bring valuable expertise beyond just capital. Having that alignment is crucial because building a company like HOPPR requires patience and long-term thinking.

Dario Priolo: Looking ahead, how do you see HOPPR's approach shaping the future of healthcare?

Dr. Khan Siddiqui: One of the most important impacts will be democratizing access to advanced AI capabilities. Today, there's a significant disparity between what's available at major academic medical centers versus community hospitals. Our platform allows smaller facilities to access the same sophisticated technology, leveling the playing field and ensuring patients get high-quality care regardless of where they live.

We're also seeing our ecosystem approach catalyze innovation across specialties. Our recent partnership with RadNet's DeepHealth to develop specialized cancer detection models is just one example of how we're expanding into new clinical areas.

But what truly excites me is how all of this technology ultimately serves our core mission: putting patients first. At the end of the day, if our technology isn't creating value for patients—improving outcomes, enhancing experiences, making care more accessible—then what's the point?

I always think in terms of 10x improvement. Can we make something 10x better, 10x faster, 10x more cost-effective? When you focus on creating that level of value for patients, everything else—revenue, growth, partnerships—follows naturally.

Dario Priolo: That's a powerful perspective. Any final thoughts you'd like to share with our audience?

Dr. Khan Siddiqui: I'd just emphasize that AI in healthcare isn't about replacing human judgment or connection—it's about enhancing it. When we relieve radiologists of repetitive tasks and help them focus their expertise where it matters most, we're not diminishing their role; we're elevating it.

The future we're building isn't one where technology replaces healthcare providers. It's one where technology helps them deliver more personalized, precise, and compassionate care. That's the true promise of AI in healthcare—making the system more human, not less.

Dario Priolo: Dr. Siddiqui, thank you for sharing these insights about HOPPR and your vision for the future of healthcare.

Dr. Khan Siddiqui: Thank you, Dario. It's been a pleasure.